Generalized Geometry–Matter Coupling Gravity

• Generalized geometry–matter coupling gravity is a family of modified theories where the gravitational action explicitly depends on both geometric and matter-sector variables.
• These models produce non-conserved energy-momentum tensors and extra force terms that alter particle motion and influence cosmological evolution.
• Different formulations, from metric to teleparallel and hybrid models, yield testable predictions such as dark-energy-like behavior and modified Newtonian limits.

Generalized geometry–matter coupling gravity denotes a family of modified gravitational theories in which the gravitational action depends not only on geometric quantities but also explicitly on matter-sector variables such as the matter Lagrangian $L_m$, the trace $T$ of the energy-momentum tensor, or related matter scalars. In these theories, the Einstein–Hilbert relation between geometry and matter is replaced by a broader interaction law, so that the coupling between curvature and matter is itself dynamical or structurally nontrivial. A recurrent consequence is that the matter energy-momentum tensor is generally not covariantly conserved, test bodies can experience an extra force orthogonal to the four-velocity, and the choice of $L_m$ ceases to be a harmless convention. This broad class includes metric $f(R,L_m)$, $f(R,T)$, and $f(R,L_m,T)$ models, teleparallel $f(T,B,L_m)$ theories, hybrid metric-Palatini models with matter-geometry coupling, and more general tensorial or auxiliary-field formulations (Haghani et al., 2021, Harko, 2014).

1. Core formulations and defining structure

A standard representative of the metric $f(R)$-type sector is the nonminimally coupled action

$$S=\int \left[f_1(R)+\bigl(1+\lambda f_2(R)\bigr)L_m\right]\sqrt{-g}\,d^4x,$$

where $f_1(R)$ and $T$0 are arbitrary functions of the Ricci scalar and $T$1 controls the coupling strength. In this setup matter is multiplied by a curvature-dependent factor, so geometry and matter are linked already at the level of the action rather than only through the field equations (Harko, 2010).

A more general curvature-matter form is

$T$2

with arbitrary functions $T$3, $T$4, and $T$5. This contains GR, ordinary $T$6 gravity, and non-minimally coupled $T$7 models as special limits, and it is one of the canonical formulations used to derive generalized energy conditions and stability criteria (Wang et al., 2012).

The unified metric theory $T$8 pushes this logic further by assuming

$T$9

so that curvature, the matter Lagrangian, and the trace $L_m$0 all enter the gravitational sector. In this sense, $L_m$1 unifies $L_m$2 and $L_m$3, while still containing ordinary $L_m$4 gravity and GR as limiting cases (Haghani et al., 2021).

Across these formulations, the energy-momentum tensor is defined by the standard metric variation of the matter action. When $L_m$5 depends only on the metric and not on its derivatives, this reduces to the familiar algebraic expression in terms of $L_m$6 and $L_m$7. What changes is not the formal definition of $L_m$8, but its dynamical role: because the action couples matter and geometry explicitly, the matter source no longer enters as an independent minimally coupled sector. This structural shift is the defining feature of generalized geometry–matter coupling gravity.

2. Matter Lagrangians, stress tensors, and non-geodesic motion

In minimally coupled GR, several perfect-fluid Lagrangians can often be used interchangeably. In generalized geometry–matter coupling gravity, that degeneracy is broken. For a perfect fluid obeying a barotropic equation of state $L_m$9, the variational analysis combined with the Newtonian limit shows that the matter Lagrangian and the corresponding stress tensor are fixed by the coupling form and by the equation of state, rather than chosen freely. In the model with action $f(R,L_m)$0, the matter Lagrangian can be written as

$f(R,L_m)$1

and the stress tensor becomes

$f(R,L_m)$2

This differs from the textbook perfect-fluid expression by an additional equation-of-state-dependent term, interpreted in the source paper as elastic/deformation energy or other internal energy content (Harko, 2010).

The same framework yields modified equations of motion. For the nonminimally coupled $f(R,L_m)$3 model, the nonconservation law is

$f(R,L_m)$4

and the corresponding particle equation of motion contains an extra force

$f(R,L_m)$5

The orthogonality condition is generic: in the unified $f(R,L_m)$6 formulation the extra force is again orthogonal to the four-velocity, so the departure from geodesic motion is a structural consequence of the coupling rather than a coordinate artifact (Harko, 2010, Haghani et al., 2021).

A central misconception addressed explicitly in the literature is that the extra force can be removed by a special choice of matter Lagrangian. For dust, $f(R,L_m)$7, the extra force does not vanish; instead it reduces to

$f(R,L_m)$8

and in the Newtonian limit the extra acceleration is

$f(R,L_m)$9

The conclusion drawn in the source analysis is that the coupling-induced force does not disappear for physical barotropic fluids and is non-zero in the dust case as well (Harko, 2010).

The cosmological dynamical-systems literature reaches a complementary conclusion. In a power-law curvature-coupling model with $f(R,T)$0, the background evolution depends crucially on whether one uses $f(R,T)$1 or $f(R,T)$2, even though these are equivalent in minimally coupled GR. In the $f(R,T)$3 case there is no extra force on particles but the energy balance is modified; in the $f(R,T)$4 case the thermodynamic behavior and phase-space structure differ. This establishes that in generalized coupling gravities the matter Lagrangian becomes part of the physical model, not merely part of its notation (An et al., 2015).

3. Cosmological dynamics and observationally constrained backgrounds

A major application of generalized geometry–matter coupling gravity is late-time cosmology. In the curvature-coupling model

$f(R,T)$5

the phase-space analysis identifies radiation-era points, a standard matter point $f(R,T)$6, and a late-time accelerated attractor $f(R,T)$7. For dust with $f(R,T)$8, the attractor has

$f(R,T)$9

so $f(R,L_m,T)$0 or $f(R,L_m,T)$1 implies acceleration, and $f(R,L_m,T)$2 approaches a de Sitter state. The same study confronts the model with Union2.1 type Ia supernova data and finds that values around $f(R,L_m,T)$3 and $f(R,L_m,T)$4 lie in the $f(R,L_m,T)$5 region (An et al., 2015).

The generalized hybrid metric-Palatini extension with non-minimal matter-geometry coupling begins from

$f(R,L_m,T)$6

and can be rewritten as a bi-scalar-tensor theory with a dynamical scalar $f(R,L_m,T)$7 and a non-dynamical matter-coupling scalar $f(R,L_m,T)$8. For $f(R,L_m,T)$9, the homogeneous FRW background recovers the standard continuity equation, although the theory remains non-conservative beyond the background level. In the explicit model $f(T,B,L_m)$0, the combined CC+Pantheon$f(T,B,L_m)$1+BAO fit yields

$f(T,B,L_m)$2

with $f(T,B,L_m)$3, and the Bayesian evidence gives $f(T,B,L_m)$4. The model reproduces an expansion history close to $f(T,B,L_m)$5CDM, places the deceleration–acceleration transition around $f(T,B,L_m)$6, and exhibits a quintessence-to-phantom transition around $f(T,B,L_m)$7 (Jalali et al., 28 Nov 2025).

A different $f(T,B,L_m)$8 cosmology uses

$f(T,B,L_m)$9

with $f(R)$0, and constructs four effective one-fluid models for stiff matter, radiation, dust, and curvature fluid. The joint analysis of 31 Cosmic Chronometer points and 1701 Pantheon+SH0ES data gives current deceleration parameters in the range

$f(R)$1

transition redshifts

$f(R)$2

and current effective equations of state

$f(R)$3

All four models show late-time transit-phase acceleration, while the stiff-fluid and radiation versions also exhibit an early accelerating phase. According to the information-criterion analysis reported there, Models III and IV are closest to $f(R)$4CDM statistically, whereas the sound-speed condition $f(R)$5 is satisfied for Models I–III but violated at late times in Model IV (Maurya et al., 16 Jan 2025).

Framework	Characteristic cosmological result	Data comparison
Power-law curvature coupling $f(R)$6	Radiation era, matter era, and stable late-time attractor $f(R)$7	Union2.1 fit with $f(R)$8 in the $f(R)$9 region
Generalized hybrid metric-Palatini with matter-geometry coupling	Background close to $$S=\int \left[f_1(R)+\bigl(1+\lambda f_2(R)\bigr)L_m\right]\sqrt{-g}\,d^4x,$$0CDM; quintessence-to-phantom transition near $$S=\int \left[f_1(R)+\bigl(1+\lambda f_2(R)\bigr)L_m\right]\sqrt{-g}\,d^4x,$$1	CC, Pantheon$$S=\int \left[f_1(R)+\bigl(1+\lambda f_2(R)\bigr)L_m\right]\sqrt{-g}\,d^4x,$$2, and DESI BAO; $$S=\int \left[f_1(R)+\bigl(1+\lambda f_2(R)\bigr)L_m\right]\sqrt{-g}\,d^4x,$$3
$$S=\int \left[f_1(R)+\bigl(1+\lambda f_2(R)\bigr)L_m\right]\sqrt{-g}\,d^4x,$$4	Late-time transit-phase acceleration in all models; early and late acceleration in Models I and II	31 CC + 1701 Pantheon+SH0ES

These results show that generalized coupling theories are not cosmologically monolithic. Some behave as effective dark-energy models with stable late-time attractors, some mimic $$S=\int \left[f_1(R)+\bigl(1+\lambda f_2(R)\bigr)L_m\right]\sqrt{-g}\,d^4x,$$5CDM closely at background level, and some produce multi-stage expansion histories. A plausible implication is that the matter-coupling sector acts as an additional model-selection layer, comparable in importance to the choice of gravitational scalar itself.

4. Thermodynamics, Newtonian limits, and viability criteria

The nonconservation of $$S=\int \left[f_1(R)+\bigl(1+\lambda f_2(R)\bigr)L_m\right]\sqrt{-g}\,d^4x,$$6 has a thermodynamic interpretation in both $$S=\int \left[f_1(R)+\bigl(1+\lambda f_2(R)\bigr)L_m\right]\sqrt{-g}\,d^4x,$$7 and $$S=\int \left[f_1(R)+\bigl(1+\lambda f_2(R)\bigr)L_m\right]\sqrt{-g}\,d^4x,$$8 gravity. Within the open-system framework, the modified balance law is rewritten as

$$S=\int \left[f_1(R)+\bigl(1+\lambda f_2(R)\bigr)L_m\right]\sqrt{-g}\,d^4x,$$9

where $f_1(R)$0 is an effective particle creation rate, and the creation pressure is

$f_1(R)$1

The entropy production is positive, the matter sector behaves as an open system, and the nonminimal coupling is interpreted as an irreversible energy transfer from geometry to matter. In this description, geometry-driven particle creation can generate large comoving entropy during cosmological evolution (Harko, 2014).

The weak-field regime can also be derived explicitly. In $f_1(R)$2 gravity, the Newtonian limit of the particle action gives a total acceleration

$f_1(R)$3

so that the coupling contributes an extra acceleration beyond the ordinary Newtonian potential. The same analysis yields a generalized Poisson equation

$f_1(R)$4

with an effective Newtonian constant determined by the local expansion of the function $f_1(R)$5 (Haghani et al., 2021).

In the Weyl–Cartan extension $f_1(R)$6, the low-velocity weak-field limit gives

$f_1(R)$7

where the effective coupling depends on the background values of $f_1(R)$8, $f_1(R)$9, $T$00, and $T$01. In that theory, nonmetricity, the Weyl vector, torsion, and the matter coupling all contribute to the renormalization of the gravitational interaction. The cosmological study in the same framework reports that the additive and multiplicative models provide good descriptions of observational data up to $T$02, and in some cases up to $T$03 (Harko et al., 2021).

Viability conditions have been formulated systematically for generalized curvature–matter coupling. For the action

$T$04

the effective attractive-gravity condition is

$T$05

and the generalized Dolgov–Kawasaki criterion is

$T$06

The same framework admits generalized SEC, NEC, DEC, and WEC inequalities written in terms of effective density and pressure, reducing to the standard GR conditions in the appropriate limit. Applied to power-law models, these conditions become explicit constraints on the parameters $T$07 and on the cosmographic quantities $T$08 (Wang et al., 2012).

5. Wormholes, inhomogeneous matter profiles, and lensing signatures

Generalized geometry–matter coupling gravity has been used extensively in wormhole physics, where the aim is to reduce or localize the exoticity normally required in GR. In $T$09 gravity with

$T$10

tideless Morris–Thorne wormholes can be sourced by Casimir energy densities. For the uncorrected Casimir models, the radial NEC is violated while the tangential NEC is satisfied; with Generalized Uncertainty Principle corrections, radial NEC violation persists, the KMM geometry is flatter, the DGS geometry is more elevated/curved, and both corrected cases satisfy the equilibrium condition $T$11 in the zero-tidal-force TOV analysis (Agrawal et al., 2024).

A linear $T$12 model,

$T$13

has been used to construct zero-tidal-force wormholes sourced by the Dekel–Zhao dark matter profile. In that setting the effective coupling enters through

$T$14

and the matter functions satisfy

$T$15

The resulting geometries satisfy the throat, flaring-out, and asymptotic-flatness conditions, allow radial NEC violation near the throat, and can accommodate both exotic and ordinary matter contributions. The lensing analysis yields a negative deflection angle, interpreted there as a repulsive gravitational effect for positive couplings (Errehymy et al., 27 May 2025).

A related $T$16 construction uses the universal density profile of cosmic voids,

$T$17

as the matter source of a traversable wormhole. For the linear model

$T$18

the field equations reduce to a first-order system for the shape function $T$19, and the strong energy condition is saturated identically through

$T$20

In the parameter study reported there, $T$21 everywhere for $T$22, the tangential NEC holds at the throat and beyond, and radial NEC violation is confined to $T$23. The deflection angle is negative throughout the explored range, so the wormhole acts as a repulsive lens (Errehymy et al., 24 Aug 2025).

Taken together, these constructions show that generalized matter couplings are being used not only to modify the effective gravitational source but also to import physically motivated matter profiles—Casimir vacuum energy, dark-matter halos, and cosmic voids—into exact or quasi-exact wormhole geometries. This suggests a shift away from ad hoc exotic fluids toward source models tied to broader astrophysical or quantum settings.

6. Alternative geometric realizations and generalized coupling prescriptions

The metric $T$24 sector is only one realization of generalized geometry–matter coupling. In teleparallel geometry, the action

$T$25

depends on the torsion scalar $T$26, the boundary term $T$27, and the matter Lagrangian. Because

$T$28

the theory contains $T$29, $T$30, and $T$31 as special cases. The flat-FLRW dynamical system is 10-dimensional in general, and specific subclasses admit de Sitter fixed points or accelerating power-law solutions (Bahamonde, 2017).

A conceptually different proposal replaces the scalar coupling constant in Einstein’s equation by a rank-4 tensor: $T$32 In this framework the vacuum sector is exactly that of GR, while the matter sector experiences a generalized tensorial coupling that can vary with observer frame, thermodynamic state, and scale. For pressureless matter, two scalar contractions $T$33 and $T$34 control the Newtonian limit through

$T$35

The source paper uses this structure to reconstruct dark-energy-like FLRW evolution and galaxy-rotation-curve-like behavior without adding new matter species (Carloni, 2016).

Another auxiliary-field realization introduces an invertible rank-2 tensor $T$36 and a physical metric

$T$37

so that matter couples to $T$38 while the pure gravitational sector remains Einstein-like. In vacuum the algebraic field equation forces $T$39, implying $T$40 and exact reduction to GR. In the Minimal Exponential Measure model, the gravitational-wave speed in matter differs from its vacuum value, and the estimate based on GW170817/GRB170817A and Earth’s mean density yields the rough bound

$T$41

The same framework admits cosmologies with an early inflationary de Sitter phase and a late-time de Sitter phase characterized by a different expansion rate (Feng et al., 2019).

The dynamical-volume-form approach modifies the measure of integration rather than the curvature sector, using a second tensor $T$42 generated from $T$43 through a fourth-order tensor $T$44,

$T$45

In the cosmological model $T$46, the key scalar variable is $T$47, and the phase portrait contains an initial accelerated era, a decelerated phase near point $T$48, and a second accelerated era associated with points $T$49 and $T$50. The late-time attractor $T$51 is realized for $T$52, with accelerated expansion for $T$53 (Boehmer et al., 2018).

Gauge-theoretic generalizations also reinterpret the coupling problem itself. In Cartan gravity, the basic gravitational variables are a contact vector $T$54 and a gauge connection $T$55, with the tetrad emerging as

$T$56

Imposing the gauge principle and polynomial simplicity yields first-order matter actions for scalar, Dirac, and Yang–Mills fields, and the usual energy-momentum and spin currents are unified into a single spin-energy-momentum three-form $T$57. The same formalism suggests an $T$58 unification of gravity with a $T$59 gauge field (Westman et al., 2012).

Generalized matter couplings also arise in multi-spin-2 theories. In massive bigravity, matter can couple consistently to effective vielbeins built from nonlinear rank-2 vielbein strings, not only to the previously known linear effective metric. The healthy class is the rank-2 construction; rank-0 and mixed constructions are generically ghostly for generic matter, while the distinctions between the new and old couplings appear only beyond the $T$60 decoupling limit or when the symmetric vielbein condition fails (Melville et al., 2015).

7. Conceptual issues, restrictions, and open problems

One persistent issue is the status of the matter Lagrangian. In minimally coupled perfect-fluid GR, $T$61 and $T$62 are often treated as interchangeable. In generalized geometry–matter coupling gravity this is no longer true: the stress tensor, force law, thermodynamic interpretation, and even the phase-space structure can depend on which $T$63 is used. The detailed analyses of nonminimally coupled $T$64 and unified $T$65 models therefore treat the matter Lagrangian as physically consequential rather than purely representational (Harko, 2010, Haghani et al., 2021).

A second issue concerns conservation. Some theories recover standard matter conservation only in restricted circumstances. In generalized hybrid metric-Palatini gravity with $T$66, the baryonic matter sector is conserved on the homogeneous FRW background, but the source paper emphasizes that this is only a background-level property and that the full theory remains intrinsically non-conservative at the perturbative level. This leaves structure growth and perturbation phenomenology as necessary next steps for assessing viability (Jalali et al., 28 Nov 2025).

A third issue is the compatibility of generalized geometry with the minimal coupling principle. In a completely general linear affine geometry, the application of MCP to Standard Model fields indicates that torsion generically creates theoretical problems: gauge-field strengths acquire torsion-dependent contributions, and fermion couplings become problematic unless the connection has no axial torsion. The conclusion drawn there is that symmetric teleparallelism, characterized by $T$67, $T$68, and nonzero nonmetricity, is the generalized affine geometry compatible with MCP, while torsionful geometries generically are not (Jimenez et al., 2020).

Several open questions are formulation-specific. In the Cartan gauge-theoretic program, the fully dynamical treatment of the contact vector $T$69, the extension of the $T$70 unification idea to non-Abelian gauge fields, and the appearance of the equivalence principle remain unresolved. In massive bigravity, the uniqueness of acceptable matter couplings depends on whether one works in the metric or vielbein language and on whether one stays below or beyond the $T$71 decoupling limit (Westman et al., 2012, Melville et al., 2015).

The field as a whole is therefore characterized by a dual movement. On one side, generalized matter couplings have been developed into a wide array of mathematically distinct frameworks—metric, teleparallel, metric-affine, Cartan, hybrid metric-Palatini, auxiliary-tensor, and multi-vielbein. On the other side, these frameworks repeatedly converge on the same physical questions: whether matter is conserved, whether free fall remains geodesic, how $T$72 should be chosen, which generalized couplings are stable, and which background-level successes survive once perturbations, local tests, and matter-sector consistency are imposed.