Lovelock Scalar-Tensor Gravity

• Lovelock scalar-tensor gravity is a framework coupling scalar fields with Lovelock densities, preserving second-order field equations and avoiding Ostrogradsky instabilities.
• It employs conformal, kinetic, and algebraic couplings to extend Brans–Dicke and Lovelock theories, with applications in cosmology, black hole physics, and inflation.
• Innovative features include scalar-tensor duality via Legendre transformation and the emergence of scalar-hairy black hole solutions that refine thermodynamic laws.

Lovelock scalar-tensor gravity refers to the broad class of gravitational theories that couple scalar fields to curvature invariants constructed from the Lovelock densities, generalizing both Brans–Dicke theory and higher-dimensional Lovelock gravity to settings with non-minimal scalar-tensor interactions. Notably, these models preserve second-order field equations for both the metric and the scalar, avoiding Ostrogradsky instabilities and absorbing much of the structure of Horndeski and Galileon-type theories in four and higher dimensions. This framework encompasses multiple methods—conformal, kinetic, and algebraic couplings—between scalar fields and dimensionally extended Euler (Lovelock) densities, with relevance for phenomenological cosmology, black hole physics, models of inflation, and the generalization of the gravitational action beyond Einstein gravity.

1. Lovelock Scalar–Tensor Lagrangians and Their Structure

The prototypical Lovelock scalar–tensor action in $D$ spacetime dimensions is

$$S = \int d^D x \sqrt{-g} \Bigg[ \sum_{k=0}^{\lfloor(D-1)/2\rfloor} \left[ a_k \, \mathcal{L}^{(k)}(g) + b_k \, \phi^{D-4k} \mathcal{S}^{(k)}(g,\phi) \right] \Bigg] + S_{\text{matter}},$$

where $\mathcal{L}^{(k)}$ are Lovelock densities of order $k$, and $\mathcal{S}^{(k)}$ are their scalar–tensor analogs obtained by replacing each Riemann tensor with a specific four-index combination built from $\phi$ and its derivatives, guaranteeing conformal invariance under $$g_{\mu\nu} \to \Omega^2 g_{\mu\nu}, \phi \to \Omega^{-1} \phi$$ (Babichev et al., 2023, Fang et al., 2022, Jiang et al., 2020).

A central example in four dimensions is Lovelock-Brans–Dicke (LBD) gravity, whose Lagrangian is

$$\mathcal{L}_{\rm LBD} = \frac{1}{16\pi} \left[\phi R + a \phi\, {}^\ast RR + b \phi\, \mathcal{G} - \frac{\omega_L}{\phi} \nabla_\alpha\phi \nabla^\alpha\phi \right]$$

where ${}^\ast RR$ is the Chern–Pontryagin density and $\mathcal{G}$ is the Gauss–Bonnet invariant (Tian et al., 2015).

Generalizations include actions with scalar-dependent couplings to arbitrary Lovelock invariants, e.g.,

$$S = \int d^D x \sqrt{-g} \left[ f_1(\phi) R + f_2(\phi)\, {}^\ast RR + f_3(\phi) \mathcal{G} - \frac{\omega(\phi)}{\phi} \nabla\phi \cdot \nabla\phi - V(\phi) \right]$$

as well as kinetic (derivative) couplings of the scalar to Lovelock tensors (Gao, 2018).

2. Field Equations and Second-Order Dynamics

The field equations derived from Lovelock scalar-tensor actions are universally second order in both the metric and the scalar, a consequence of the antisymmetrization and structural properties of Lovelock densities. Varying with respect to $g_{\mu\nu}$ and $\phi$ gives, e.g.,

$$\sum_{k=0}^{k_{\text{max}}} a_k \mathcal{G}^{(k)}_{\mu\nu} = \sum_{k=0}^{k_{\text{max}}} b_k T^{(k)}_{\mu\nu} + 8\pi T^{\text{matter}}_{\mu\nu},$$

where $\mathcal{G}^{(k)}_{\mu\nu}$ are metric variations of $\mathcal{L}^{(k)}$, and $T^{(k)}_{\mu\nu}$ are stress tensors from the scalar-coupled terms. The scalar equation has the generic form

$$\sum_{k=0}^{k_{\text{max}}} \left[ b_k (D-4k)\, \phi^{D-4k-1} \mathcal{S}^{(k)} + b_k \phi^{D-4k} \nabla \cdots (\text{terms}) \right] = 0$$

(Fang et al., 2022, Tian et al., 2015, Babichev et al., 2023, Jiang et al., 2020).

The crucial property is the absence of higher-order derivatives in the final form of the equations, due to generalized Bianchi and conformal identities. This property holds also for kinetic couplings via Lovelock tensors, where the scalar field equation becomes

$$\sum_p \beta_p G^{(p)}_{\mu\nu} \nabla^\mu \nabla^\nu \phi + V_{,\phi} = 0,$$

with divergence-free $G^{(p)}_{\mu\nu}$ (Gao, 2018).

3. Scalar-Tensor Duality and Relation to $f($Lovelock$)$ and Horndeski Theories

Any $f($Lovelock$)$ theory, i.e., with a Lagrangian $f$ of the Lovelock invariants, admits an equivalent scalar-tensor description:

$$S = \int d^D x \sqrt{-g} \left[ \sum_{k=0}^{K} \phi_k \mathcal{L}_k - V(\phi_0, ..., \phi_K) \right],$$

with auxiliary scalars $\phi_k = \partial f / \partial \chi_k$ (Legendre dualization), potential $V$ the associated Legendre transform, and field equations with second derivatives only (Bueno et al., 2016, Lachaume, 2017). For $f(R)$ (single density), this reproduces Brans–Dicke gravity ($\omega=0$).

The Legendre transform is invertible if the Hessian of $f$ is non-degenerate; otherwise, only a subset of the scalars are dynamical. In $D=4$, the effective scalar-tensor model sits inside the Horndeski class (generalized Galileon), ensuring absence of Ostrogradsky ghosts. When the Hessian vanishes (pure Lovelock theory), no extra scalars propagate—the metric field equations revert to those of Lovelock gravity (Bueno et al., 2016, Tian et al., 2015, Charmousis, 2014, Kobayashi, 2020).

Upon Kaluza–Klein reduction from higher dimensions, Lovelock gravity yields Horndeski-like scalar-tensor actions in $D=4$, with full ghost-free structure preserved (Charmousis, 2014, Kobayashi, 2020).

4. Black Hole Solutions, Scalar Hair, and Thermodynamics

A distinguishing feature is the existence of scalar-hairy black hole solutions, frequently exhibiting "stealth" character (scalar field non-trivial, but with vanishing energy-momentum tensor, so the metric solves the pure Lovelock field equations) (Babichev et al., 2023, Fang et al., 2022). In these backgrounds, static as well as time-dependent scalar field configurations are allowed even in the absence of shift symmetry.

The thermodynamic analysis reveals that under the dominant energy condition, the surface gravity $\kappa$ is constant over Killing horizons (zeroth law) (Fang et al., 2022). Entropy functionals for these theories display notable subtleties: the Wald entropy formula, which extends the area law to general higher curvature gravity, may fail the linearized second law in scalar-hairy Lovelock models with genuine higher-order terms. Instead, the conformally-related Jacobson–Myers (JM) entropy agrees with the second law at linear order in processes such as Vaidya-like collapse, with additional scalar contributions beyond those of $F$(Riemann) gravity (Jiang et al., 2020).

The entropy, Killing horizon properties, and the behavior of scalar hair are controlled by the conformal structure of the theory and the non-minimal couplings.

5. Cosmological Implications and Modified Gravity Phenomenology

Lovelock scalar-tensor gravities are central to several cosmological scenarios:

• In higher dimensions ($n \geq 7$), kinetic couplings to Lovelock tensors induce cyclic, eternally expanding/contracting, or multistage de Sitter cosmologies due to the richer structure of higher-order invariants (Gao, 2018).
• In $D=4$, regularized Lovelock gravity gives rise, after Kaluza–Klein reduction and appropriate rescalings, to an effective Horndeski scalar-tensor theory where scalar charges source radiation-like components decaying as $a^{-6}$, which can influence early-universe cosmology or be screened in local gravity (Kobayashi, 2020).
• In inflationary phenomenology, quadratic $f$(Lovelock) theories such as $f(L) = L + \beta L^2$, with $L = R + \alpha \mathcal{G}/4$, produce scalar-tensor models that, in the Einstein frame, yield Starobinsky-like inflationary potentials modified by Gauss–Bonnet-coupled Galileon terms. This scenario aligns with recent CMB data when the Gauss–Bonnet coupling is tuned, and is algebraically equivalent to Higgs inflation with a Gauss–Bonnet term (Addazi et al., 24 Dec 2025).

6. Uniqueness, Mathematical Structure, and Open Directions

Lovelock’s theorem guarantees that the Lovelock densities furnish the unique scalar invariants built from the metric giving second-order field equations in higher dimensions. Scalar-tensor extensions—either algebraic or via kinetic (derivative) coupling—preserve this property. The equivalence to $f$(Lovelock) via Legendre transformation reveals all degree-of-freedom content and ensures that, except in degenerate cases, models propagate the massless graviton together with zero or more scalars, but no ghostly higher-derivative modes (Bueno et al., 2016, Lachaume, 2017).

Key open problems include:

• Nonlinear and quantum generalizations of the second law of black hole thermodynamics for these models (Jiang et al., 2020).
• Extension of the black hole rigidity theorem to the higher-curvature scalar-tensor context (Fang et al., 2022).
• Full classification of stealth versus non-stealth scalar configurations, especially when conformal invariance is broken (Babichev et al., 2023).
• Systematic study of screening mechanisms and local gravity constraints in the presence of nontrivial Lovelock–scalar couplings (Kobayashi, 2020).

Overall, Lovelock scalar-tensor gravity provides a technically robust, physically rich extension of both scalar-tensor and higher-curvature gravity theories, with implications spanning from early-universe inflation, black hole thermodynamics, to self-tuning and cosmological constant problems. Its mathematical structure underpins and links much of the broader theory space of modified gravity (Tian et al., 2015, Babichev et al., 2023, Addazi et al., 24 Dec 2025).