Hybrid Metric-Palatini Gravity Theory

• Hybrid metric-Palatini gravity is a modified theory that integrates the Einstein-Hilbert term with a Palatini correction, introducing a light, dynamic scalar degree of freedom.
• The theory recasts into a scalar-tensor framework, enabling applications from cosmic acceleration and galactic rotation curves to modified black hole and stellar structures.
• Analyses in weak-field, cosmological, and astrophysical regimes demonstrate its consistency with solar-system constraints and its potential to address dark matter and dark energy phenomena.

Searching arXiv for the core review and representative papers on hybrid metric-Palatini gravity. Hybrid metric-Palatini gravity is a modified gravity theory designed to combine the best features of the metric and Palatini formalisms used in $f(R)$ gravity. In its original form, it keeps the Einstein-Hilbert term built from the metric Ricci scalar $R$ and adds a Palatini correction $f(\mathcal R)$, where $\mathcal R$ is constructed from an independent connection. A generalized extension takes the gravitational Lagrangian to be a function $f(R,\mathcal R)$. Across the literature, the framework is repeatedly described as admitting a genuinely dynamical scalar degree of freedom that can be light and long-ranged, can pass local gravity tests, and can modify galactic and cosmological dynamics, including late-time cosmic acceleration (Capozziello et al., 2015, Harko et al., 2020, Tamanini et al., 2013, Capozziello et al., 2013).

1. Origins, motivation, and defining action

The theory was introduced to bridge two standard variational approaches that cease to be equivalent once one leaves Einstein gravity. In metric $f(R)$ gravity, the metric is the only dynamical field, the extra scalar degree of freedom is dynamical, and Solar System viability typically forces that scalar to be very massive unless screening is invoked. In Palatini $f(R)$ gravity, the metric and connection are independent and the field equations are second order, but the scalar degree of freedom is non-dynamical and algebraically tied to matter, which leads to pathologies such as microscopic instabilities, surface singularities, and infinite tidal forces in some models. The hybrid theory was proposed precisely to retain the Einstein-Hilbert term in metric form while adding a Palatini-type correction (Capozziello et al., 2015, Harko et al., 2020, Bronnikov, 2019).

The basic hybrid action is written as

$$S=\frac{1}{2\kappa^2}\int d^4x \sqrt{-g}\,\big[R+f(\mathcal R)\big]+S_m,$$

where $R$ is the usual metric Ricci scalar, $\mathcal R=g^{\mu\nu}\mathcal R_{\mu\nu}$ is the Palatini curvature scalar built from an independent connection $R$0, and $R$1. The theory is “hybrid” because it contains metric curvature in the Einstein-Hilbert term and Palatini curvature in the modification (Capozziello et al., 2015, Harko et al., 2020).

The generalized hybrid theory extends this construction to

$R$2

placing the metric and Palatini curvature scalars on equal footing at the level of the action. This broader setup opens the door to richer scalar-tensor dynamics, including two interacting scalar fields (Tamanini et al., 2013, Rosa, 2019).

2. Scalar-tensor structure and dynamical degrees of freedom

A central result of the subject is that hybrid metric-Palatini gravity can be rewritten as a scalar-tensor theory. Introducing an auxiliary scalar field $R$3, one obtains

$R$4

Variation with respect to the independent connection gives

$R$5

so the Palatini connection becomes the Levi-Civita connection of the conformally related metric

$R$6

Eliminating the independent connection yields the scalar-tensor action

$R$7

This resembles a Brans-Dicke theory with $R$8, but the literature stresses that it is not the same as the usual Brans-Dicke theory in the way the scalar couples to curvature, and that the scalar is dynamical rather than algebraically constrained (Capozziello et al., 2015, Harko et al., 2020, Capozziello et al., 2012).

The metric field equations can be written as

$R$9

while the scalar obeys the Klein-Gordon-type equation

$f(\mathcal R)$0

The scalar is therefore propagating and sourced by the matter trace $f(\mathcal R)$1 (Capozziello et al., 2015, Harko et al., 2020).

For the generalized theory $f(\mathcal R)$2, the scalar-tensor representation becomes bi-scalar. Defining

$f(\mathcal R)$3

the action can be rewritten as

$f(\mathcal R)$4

This exposes one scalar that couples directly through $f(\mathcal R)$5 and a second scalar with non-standard kinetic structure (Tamanini et al., 2013, Rosa, 2019, Gomes et al., 15 Jun 2025).

A further generalization introduces non-metricity and generalized conformal transformations. In that framework, metric, Palatini, and hybrid metric-Palatini theories sit inside a broader scalar-tensor class with arbitrary functions $f(\mathcal R)$6. The key result is that the generalized hybrid theory is on-shell equivalent to a purely metric scalar-tensor theory with the same metric and scalar-field solutions, while the affine connection is reconstructed from the metric frame (Borowiec et al., 2020).

3. Mathematical consistency and weak-field viability

The Cauchy problem has been studied in detail in the scalar-tensor formulation. Using generalized harmonic coordinates,

$f(\mathcal R)$7

the field equations can be rewritten as a quasi-diagonal, quasi-linear second-order system,

$f(\mathcal R)$8

together with the scalar field equation. The theory is therefore well-formulated, and it is well-posed in suitable Sobolev spaces provided the initial data satisfy the constraints and gauge conditions. The positive result extends to standard matter sources, including electromagnetic fields, Yang-Mills fields, perfect fluids, dust, and Klein-Gordon scalar fields, under the standard conservation law $f(\mathcal R)$9 (Capozziello et al., 2013).

In the weak-field limit, one expands around a background $\mathcal R$0. The scalar perturbation satisfies a Yukawa-type equation,

$\mathcal R$1

with

$\mathcal R$2

One expression used in stellar and vacuum analyses writes the effective Newtonian coupling and PPN parameter as

$\mathcal R$3

Solar-system compatibility can then be obtained either by making the scalar heavy or by requiring $\mathcal R$4, so that the Yukawa corrections remain small even if the scalar is long-ranged (Danila et al., 2016, Danila et al., 2018).

This local behavior is one of the defining claims of the theory. The review literature emphasizes that hybrid metric-Palatini gravity can pass local tests even when the scalar field is very light, and that the original hybrid model was presented as unifying cosmic acceleration with solar-system constraints without resource to the chameleon mechanism (Capozziello et al., 2015, Rosa, 2019).

4. Cosmology and generalized cosmological frameworks

For a spatially flat FLRW universe,

$\mathcal R$5

the modified Friedmann equations in the scalar-tensor representation can be written as

$\mathcal R$6

$\mathcal R$7

and

$\mathcal R$8

The literature summarizes de Sitter expansion, marginally accelerating solutions, power-law acceleration, and cosmological phase transitions within this framework (Harko et al., 2020).

The Einstein static universe has also been analyzed in the dynamically equivalent scalar-tensor description. Stable regions are parametrized by the first and second derivatives of the scalar potential. An important result is that, unlike in GR, an Einstein static solution can exist for negative curvature $\mathcal R$9, provided $f(R,\mathcal R)$0. In the closed case, however, the stability regions for homogeneous and inhomogeneous scalar perturbations do not overlap (Boehmer et al., 2013).

The generalized $f(R,\mathcal R)$1 theory has been studied by reconstruction methods and dynamical systems. Exact FLRW solutions include de Sitter expansion, radiation-like vacuum expansion, collapsing universes, and matter-like power-law behavior in vacuum. At the phase-space level, the generalized theory has no global attractors; stable universes can either be described by scale factors that diverge in finite time or asymptotically approach constant values. The same literature concludes that higher-order terms of order six and eight in derivatives of the metric are not neglectable (Tamanini et al., 2013, Rosa, 2019).

Several later extensions make the cosmological sector more structured. A class of scalar-tensor theories with non-metricity unifies metric, Palatini, and hybrid actions through generalized conformal transformations and generalized invariants; in the Starobinsky example $f(R,\mathcal R)$2, the hybrid formulation is quantitatively different from the metric one, and the hybrid Starobinsky model yields $f(R,\mathcal R)$3, which is strongly disfavored (Borowiec et al., 2020). A non-local hybrid extension shows that generic non-local models are ghostly, but carefully chosen “split” hybrid constructions can be ghost-free and support inflation in an Einstein-frame multi-field scenario with a Starobinsky-like plateau and spectator-field behavior (Bombacigno et al., 2024).

An even more recent extension considers

$f(R,\mathcal R)$4

which cannot in general be analytically transformed to a scalar-tensor theory. On FRW, however, a metric-compatible connection can still be solved for. The model fits background expansion almost as well as $f(R,\mathcal R)$5CDM, with

$f(R,\mathcal R)$6

while differing from $f(R,\mathcal R)$7CDM in the higher-order diagnostics $f(R,\mathcal R)$8, $f(R,\mathcal R)$9, $f(R)$0, in the statefinder plane, and in the $f(R)$1 diagnostic (Shahidi et al., 30 Jun 2025).

5. Galactic, stellar, cluster, and accretion phenomenology

At galactic scales, the scalar sector has been used as an effective geometric halo. For static spherical systems, circular orbits satisfy

$f(R)$2

and in the weak-field halo the scalar profile is Yukawa-like,

$f(R)$3

In the intermediate region $f(R)$4, the model yields an approximately constant tangential velocity. The rotation-curve analysis explicitly compared the theory with the low surface brightness galaxies DDO189, UGC1281, UGC711, and UGC10310, using the parameter choice

$f(R)$5

and reported good agreement with the observed rotation curves (Capozziello et al., 2013).

At cluster scales, the review literature describes a generalized virial theorem in which the scalar field contributes a geometric mass term. In that picture, the virial mass discrepancy can be attributed to an effective geometric mass $f(R)$6, rather than to particle dark matter (Capozziello et al., 2015, Harko et al., 2020).

Compact-star structure is modified through hybrid analogues of the mass continuity and Tolman-Oppenheimer-Volkoff equations. Numerical studies with stiff-fluid, radiation-like, MIT bag-model, and Bose-Einstein-condensate equations of state found that, for all the considered equations of state, hybrid metric-Palatini gravity stars are more massive than their general relativistic counterparts. The paper also noted that a constant scalar field forces the matter equation of state into a bag-model form describing quark matter (Danila et al., 2016).

Thin accretion disks around static spherically symmetric black holes have been studied with the Novikov-Thorne model. The central result is that accretion disks in hybrid metric-Palatini gravity are generally colder and less luminous than in general relativity, although realistic solar-system-constrained parameter choices keep the observables very close to Schwarzschild (Dyadina et al., 2023).

System	Main result	Source
Galactic rotation curves	Flat rotation curves can be reproduced by the scalar sector	(Capozziello et al., 2013)
Galaxy clusters	Virial mass discrepancy can be explained by geometric mass	(Capozziello et al., 2015)
Hybrid stars	Stars are more massive than their GR counterparts	(Danila et al., 2016)
Thin accretion disks	Disks are generally colder and less luminous than in GR	(Dyadina et al., 2023)

6. Exact spacetimes, causality, perturbations, and higher-dimensional realizations

Hybrid metric-Palatini gravity supports several nonstandard exact or numerical spacetime sectors. Wormhole geometries were among the earliest explicit applications. In the scalar-tensor representation, the flare-out condition is imposed on the effective stress-energy tensor,

$f(R)$7

so ordinary matter threading the wormhole need not violate the null energy condition. Two explicit solutions were presented: one non-asymptotically flat geometry requiring matching to an exterior vacuum region, and one asymptotically flat geometry for which

$f(R)$8

so the NEC at the throat is satisfied if $f(R)$9 (Capozziello et al., 2012).

The vacuum spherical sector has both numerical and analytic branches. Numerical studies of static, spherically symmetric vacuum solutions detected black hole formation from singular behavior in the metric functions, with horizon location and thermodynamic properties depending on the scalar background and on the scalar potential (Danila et al., 2018). In the special massless vacuum sector $f(R)$0, $f(R)$1, $f(R)$2, the theory becomes mathematically equivalent to GR with a phantom conformally coupled scalar field. In that sector, generic asymptotically flat solutions either contain naked central singularities or describe traversable wormholes, while a special two-parameter family yields globally regular black hole solutions with extremal horizons; those black hole and wormhole solutions are argued to be unstable under monopole perturbations (Bronnikov, 2019).

The theory also admits rotating homogeneous spacetimes of Gödel type in the presence of non-minimal matter coupling through $f(R)$3. With a perfect-fluid source, the field equations imply

$f(R)$4

which is the Gödel metric and therefore non-causal. With a scalar field, or with a perfect fluid plus scalar field, they imply

$f(R)$5

for which the critical radius goes to infinity and the solution is causal (Gonçalves et al., 2021).

In gravitational-wave propagation, the generalized scalar-tensor representation of hybrid metric-Palatini gravity yields tensor perturbations that propagate as massless waves at the speed of light and carry the usual $f(R)$6 and $f(R)$7 polarizations. The scalar sector produces two additional scalar modes that are generally massive and therefore subluminal, together with breathing and longitudinal distortions in the Newman-Penrose analysis. Because the scalar masses depend solely on the interaction potential, the potential can be fine-tuned so that the scalar modes become massless and luminal; the gravitational-wave study therefore argues that the theory may be effectively unfalsifiable from gravitational-wave propagation alone (Gomes et al., 15 Jun 2025).

Higher-dimensional realizations are also part of the subject. In five dimensions, the hybrid metric-Palatini brane system admits thick branes generated either by a background scalar field or by pure gravity. In both models, the branes are single thick branes with no inner structure in the energy density, the tensor sector reduces to a supersymmetric Schrödinger problem with

$f(R)$8

the graviton zero mode is normalizable, and effective four-dimensional gravity is reproduced on the brane (Fu et al., 2016).

Generalized junction conditions have been derived for $f(R)$9 gravity in both geometric and scalar-tensor form. One explicit consequence is that a Minkowski interior matched to a Schwarzschild exterior cannot be joined at arbitrary radius: the shell must lie at

$$S=\frac{1}{2\kappa^2}\int d^4x \sqrt{-g}\,\big[R+f(\mathcal R)\big]+S_m,$$0

the Buchdahl radius in GR. The same formalism yields a quasistar with a black hole and smooth matching at the light ring, and a wormhole with a Schwarzschild-AdS exterior for which the null energy condition is obeyed (Rosa et al., 2021).