Reworking Brans-Dicke Gravity: A New Route to Dark Energy and the Hubble Tension

For more than two decades, the standard model of cosmology, known as Lambda-Cold Dark Matter or LambdaCDM, has been remarkably successful at matching almost every observation we can throw at it. Yet cracks are appearing. The most famous is the Hubble tension, a stubborn disagreement between the expansion rate of the universe measured in the early cosmos and that measured nearby. There are also hints of tension in how fast large-scale structures grow. These anomalies have pushed theorists to explore alternatives in which dark energy is not a static cosmological constant but a dynamical component that evolves over cosmic time. A new study published in General Relativity and Gravitation by Matthew Debono, Jackson Levi Said and Gabriel Farrugia, researchers at the University of Malta and MCAST, examines one such alternative in meticulous mathematical detail: a generalised, non-canonical version of Brans-Dicke theory.

Brans-Dicke theory, first proposed in the 1960s, is one of the oldest and most celebrated extensions of Einstein’s general relativity. In it, the strength of gravity itself is not fixed but governed by a scalar field that varies across space and time. The original theory contains a single adjustable parameter, the Brans-Dicke parameter, which controls how strongly this scalar field interacts with the curvature of spacetime. Solar system experiments, notably radio tracking of the Cassini spacecraft, constrain this parameter to be enormous, roughly greater than ten thousand, which pushes ordinary Brans-Dicke theory very close to general relativity and limits its usefulness as a cosmological alternative. The new work revives the framework by generalising it in two directions at once: the scalar field is non-minimally coupled to gravity through a function of the field, and its kinetic term is non-canonical, meaning the field’s motion contributes to the cosmic energy budget in a field-dependent way rather than in the standard quadratic form.

Technically, the authors work with an action in which the Ricci scalar is multiplied by a function f of the scalar field, taken to be one plus a coupling constant xi times the field, while the scalar sector takes the form of a kinetic term divided by that same function, minus a potential. This particular structure is inspired by Noether symmetry arguments, a mathematical technique for selecting physically consistent Lagrangians. The construction is deliberately broad: setting the coupling xi to zero and choosing a constant potential recovers standard LambdaCDM, while other choices reproduce quintessence and the original Brans-Dicke theory. The non-canonical kinetic term, in which the usual kinetic energy is scaled by the inverse of the gravitational coupling function, is what distinguishes this model from most scalar-tensor studies and, as the authors show, noticeably changes the resulting cosmic dynamics even when the dynamical variables are defined in ways similar to those used in simpler models.

To analyse the cosmology, the team assumes a flat, homogeneous and isotropic universe filled with pressureless matter and radiation, then rewrites the modified Friedmann equations and the scalar field equation as an autonomous system of first-order differential equations. This dynamical systems approach, a staple of modern theoretical cosmology, replaces the difficult task of solving the equations outright with the more tractable problem of finding equilibrium states, called critical points, and determining whether they are stable. The dimensionless variables capture the kinetic and potential energy of the scalar field relative to the total energy density, along with rescaled matter and radiation densities. Because of the non-minimal coupling, these rescaled densities are not the physical ones, but they relate to them through the coupling function, a subtlety the authors handle carefully throughout.

A key constraint shapes the whole analysis. For the phase space of solutions to be bounded, and therefore mathematically well-behaved, the coupling parameter omega must be positive; if it were negative, the allowed region of the phase space would become a hyperbola stretching to infinity, and the phantom-like field it describes would require more elaborate compactification techniques. Positivity of omega also guarantees the absence of ghosts and Laplacian instabilities, the pathological modes that render many modified gravity theories unphysical. With these restrictions in place, the authors investigate three widely used potentials in turn: a constant potential, a power-law potential and an exponential potential, visualising the flow of cosmic histories with bounded phase portraits.

For the constant potential, the analysis uncovers a sequence of critical points that mirrors the broad architecture of LambdaCDM. A radiation-dominated saddle point gives way to a mixed matter-and-scalar-field epoch, and the universe finally settles into a scalar-field-dominated attractor that can drive accelerated expansion, provided omega exceeds half of xi squared, a condition easily satisfied when deviations from general relativity are small. Intriguingly, the matter-dominated point in this model always carries some scalar field contribution, a feature absent from simpler quintessence models, though tuning the couplings can suppress it. Numerical evolutions show that choosing omega at least an order of magnitude larger than xi shifts matter-radiation equality to a redshift of order one thousand, matching observations, whereas smaller values of omega push equality to redshifts around one hundred, which is ruled out by data.

The power-law potential, written in a form that keeps the key dynamical variable constant, produces richer behaviour with two additional saddle points at which the scalar field mixes with radiation or matter. These mixed epochs can even exhibit transient acceleration for a narrow range of negative exponents, though the authors caution that this does not guarantee a consistent later cosmology. For the illustrative inverse-square case, viability requires the matter-dominated point to be a saddle and the late-time attractor to exist, which constrains the ratio of omega to xi squared to lie either below one half or above six and a half. Notably, the team finds no tracking behaviour of the kind familiar from minimally coupled quintessence, which they attribute to the non-minimal coupling, since tracking solutions were never designed for such theories. They also verify that their critical points reduce exactly to those of standard Brans-Dicke cosmology under an appropriate limit, a reassuring consistency check.

The exponential potential brings the most surprising result. Because of the non-minimal coupling, the slope variable of the potential becomes dynamical rather than fixed, and this generates an entirely new critical point: a fully scalar-field-dominated, dark and accelerating attractor that cannot exist in standard quintessence with an exponential potential. This point is stable for all relevant values of the coupling and serves as a genuine late-time dark energy attractor. However, the model also exhibits an early epoch of kination, in which the kinetic energy of the scalar field dominates, which does not appear in the standard cosmic history and must be suppressed through careful choice of initial conditions. Fitting the physical density parameters to a LambdaCDM-like evolution suggests that couplings of order one require an essentially flat potential, while smaller couplings allow a gentle exponential slope of order one tenth.

Across all three potentials, a consistent picture emerges: viable cosmologies exist, but they demand that deviations from general relativity be modest, with the coupling xi comfortably below one, and that the parameters satisfy specific stability and acceleration conditions summarised by the authors in a compact table. The dynamical behaviour differs noticeably from other scalar-tensor models despite superficially similar variables, underlining how the combination of non-minimal coupling and non-canonical kinetics reshapes the cosmic phase portrait. None of the models provides an early inflationary starting point from the dynamics alone, meaning these scenarios are best understood as descriptions of the late universe, though the exponential case offers a saddle point that could act as an inflationary transition for suitable parameters.

The study is a background-level analysis, and the authors are explicit about what comes next. Cosmological perturbations in these models must be evolved and confronted with observations of structure growth and the cosmic microwave background before any claim can be made that they ease the Hubble tension or the clustering tension. The team also flags the Palatini formalism, an alternative variational approach in which Brans-Dicke parameters can take much lower values, as a promising direction. Still, the message of the work is striking: a sixty-year-old alternative to Einstein’s gravity, generalised with a non-canonical twist, can reproduce the entire background history of our universe, from radiation domination through matter domination to today’s accelerated expansion, and may yet hold clues to the anomalies that currently trouble precision cosmology.

Subject of Research: Cosmological dynamics of a non-canonical generalised Brans-Dicke scalar-tensor theory as an alternative to LambdaCDM dark energy

Article Title: Cosmological dynamics of a non-canonical generalised Brans-Dicke theory

Article References: Debono, M., Said, J. L., & Farrugia, G. (2026). Cosmological dynamics of a non-canonical generalised Brans-Dicke theory. General Relativity and Gravitation, 58(10), Article 116. https://doi.org/10.1007/s10714-026-03616-0

Image Credits: AI Generated

DOI: 10.1007/s10714-026-03616-0

Keywords: Brans-Dicke theory, dark energy, cosmology, dynamical systems, scalar-tensor gravity, Hubble tension, LambdaCDM, modified gravity, quintessence, critical points, non-canonical scalar field, General Relativity and Gravitation