How to Glue Two Different Theories of Gravity Together

What happens at the seam where one theory of gravity ends and another begins? It sounds like a philosophical riddle, but for physicists constructing exotic spacetimes — black holes without singularities, universes that switch gravitational laws during a cosmic phase transition, or models that hide modified gravity from laboratory tests — it is a hard technical question with real consequences. A new theoretical study by Amin Aalipour and Nima Khosravi of Sharif University of Technology, published in The European Physical Journal C, provides a systematic answer for one of the most popular classes of modified gravity theories: the so-called f(R) family, in which the usual Einstein-Hilbert action built from the Ricci scalar R is replaced by an arbitrary function f(R) of that same curvature.

The mathematical rules for sewing two spacetimes together are known as junction conditions, and they have a storied history. In 1966, Werner Israel worked out the matching rules for two regions described by ordinary Einstein gravity separated by a thin shell of matter — the framework that underlies everything from wormhole physics to membrane models of the universe. But f(R) gravity is a different beast. Because the action contains higher derivatives of the metric, the naive generalization of Israel’s conditions is far from obvious, and the literature has produced subtly conflicting answers. The new work settles the question using a variational approach, deriving the bulk field equations and the boundary matching rules simultaneously from a single well-defined action principle.

The setup is geometrically clean. The authors consider two smooth, orientable four-dimensional manifolds, each governed by its own gravitational function — call them f-plus applied to the curvature R-plus on one side, and f-minus applied to R-minus on the other. These regions are joined along a non-null hypersurface, a three-dimensional boundary that can be timelike or spacelike depending on the sign convention for its normal vector. The most basic requirement, inherited from demanding that coordinates and the metric be well defined at the seam, is the Lichnerowicz condition: the full metric and its induced boundary metric must be continuous across the interface. From there, the real content of the matching problem lies in how the extrinsic curvature — a measure of how the boundary bends within the surrounding spacetime — behaves on each side.

To see what the theory demands, the authors first warm up with Einstein gravity, but with a twist: they allow the two sides to have different gravitational couplings, effectively different Newton’s constants, kappa-plus and kappa-minus. Varying the full action, including the Gibbons-Hawking-York boundary term that makes the variational problem well posed for fixed boundary metrics, they recover a generalized Israel condition. The jump in the combination of extrinsic curvature terms, weighted by the appropriate coupling on each side, must be balanced by a surface energy-momentum tensor living on the boundary. When the couplings are equal, the familiar textbook result drops out exactly — a reassuring consistency check on the machinery.

The heart of the paper is the f(R) case. The authors exploit a classic trick: any f(R) theory can be rewritten as a scalar-tensor theory by introducing an auxiliary field Phi, defined as the derivative of f with respect to R. This field, which carries the extra scalar degree of freedom that f(R) gravity adds to general relativity, obeys its own dynamical equation — a Klein-Gordon-type relation connecting the Laplacian of Phi, the Ricci scalar, and the function f itself. When the action, including its carefully constructed boundary term proportional to Phi times the trace of the extrinsic curvature, is varied, the junction conditions emerge with striking clarity. The auxiliary field Phi must be continuous across the interface, its tangential derivatives must match, and the trace of the extrinsic curvature must be continuous. The tensorial matching rule then takes a compact form: the jump in the traceless part of the extrinsic curvature, multiplied by the common value of Phi, plus a term involving the jump in the normal derivative of Phi, must be balanced by any surface stress-energy on the shell.

Here is where the result turns genuinely surprising. Earlier analyses, using geometric and distributional methods, had concluded that the Ricci scalar R itself must be continuous across any junction in f(R) gravity, even when a thin shell of matter is present. The variational approach tells a different story. What must be continuous is not R but the derivative of f with respect to R — the quantity Phi. For a single, smooth, invertible function f shared by both sides, continuity of Phi does force continuity of R, recovering the earlier results. But when the two sides host genuinely different gravitational functions, f-plus and f-minus, the continuity of Phi only imposes a relation between R-plus and R-minus at the boundary, not their equality. The scalar curvature can jump. In other words, two different theories of gravity can be glued together consistently even though the curvature itself is discontinuous at the seam — a possibility invisible in all previous studies, which had implicitly assumed the same f on both sides.

The authors go further and check that this picture survives the change of viewpoint that theorists love most: the conformal transformation to the Einstein frame. In the Jordan frame, where matter couples directly to the metric g, the matching is expressed through Phi. In the Einstein frame, the same physics is described by an ordinary Einstein-Hilbert action plus a scalar field phi with a potential, nonminimally related to the original metric through a conformal factor Omega. The transformation is delicate — boundary terms do not transform trivially under rescalings of the metric — but the authors show, both by transforming the variation of the action and by transforming the action itself, that the two frames yield identical junction conditions. Continuity of the Einstein-frame scalar field phi implies continuity of Omega, which implies continuity of Phi, and the extrinsic curvature conditions map onto one another exactly. The equivalence between frames, often taken for granted in the bulk, thus extends to the subtle business of gluing.

Why should anyone care about sewing mismatched gravities together? The authors point to several concrete arenas. In screening mechanisms such as the symmetron model, the strength of the extra gravitational force depends on the ambient matter density, so the transition between a screened, Einstein-like region and an unscreened, modified-gravity region can be extremely sharp — precisely the situation where junction conditions matter. In uber-gravity, a proposal in which different gravitational theories are stitched together at a fixed value of the scalar curvature to address the cosmological constant problem, the transition is sharp by construction. And in regular black hole models, where a modified-gravity core replaces the singular center of a classical black hole, one needs to know exactly how the interior and exterior geometries must communicate across their shared boundary. The new conditions supply the mathematical glue for all of these constructions.

The technical payoff is a clean set of smooth-junction rules. When no thin shell of matter is present, matching two different f(R) theories requires continuity of Phi, equality of the full extrinsic curvature tensors on the two sides, and equality of the full gradient of Phi — normal and tangential components alike. Notably, even a perfectly smooth transition between different gravitational models generally leaves the Ricci scalar discontinuous, because the relation between Phi and R differs on each side of the seam. Only when the two functions coincide does the matching force R-plus equals R-minus, collapsing the new conditions onto the standard ones found in the literature.

The work leaves open several extensions that the authors flag for the future: junctions across null hypersurfaces, where the boundary geometry degenerates and the boundary terms must be modified; dynamical boundary Lagrangians with their own gravitational degrees of freedom; and comparisons with distributional treatments of higher-derivative theories. But the central message is already clear and, for a field that has debated these conditions for nearly two decades, quietly revolutionary. The universal quantity that must pass smoothly through a gravitational seam is not curvature itself but the scalar degree of freedom that f(R) gravity smuggles into Einstein’s theory. Get that field and the bending of the boundary right, and you can glue almost any two gravitational worlds together — with or without a thin shell of matter holding the joint in place.

Subject of Research: Junction conditions for matching different f(R) modified gravity theories across a hypersurface

Article Title: Gluing different gravitational models: f(R) case

Article References: Aalipour, A., & Khosravi, N. (2026). Gluing different gravitational models: f(R) case. The European Physical Journal C, 86(9), Article 1130. https://doi.org/10.1140/epjc/s10052-026-16336-1

Image Credits: AI Generated

DOI: 10.1140/epjc/s10052-026-16336-1

Keywords: f(R) gravity, junction conditions, modified gravity, general relativity, extrinsic curvature, Jordan frame, Einstein frame, conformal transformation, scalar-tensor theory, black holes, screening mechanisms, theoretical physics