⚛️ general relativity

Removability criterion for radiative corrections to the Starobinsky attractor in weakly nonlocal gravity

This paper establishes a removability criterion for radiative corrections in weakly nonlocal gravity, demonstrating that while certain one-loop effects preserve the Starobinsky attractor's predictions by evolving the model into the E-family, other deformations (such as specific $R^n$ terms or matter-induced corrections) fail this test and significantly shift the inflationary observables.

Original authors: Ekapob Kulchoakrungsun, Phongpichit Channuie, Daris Samart

Published 2026-08-19

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Ekapob Kulchoakrungsun, Phongpichit Channuie, Daris Samart

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the earliest moments of the universe, a fraction of a second after the Big Bang, space itself is believed to have undergone a period of explosive expansion known as inflation. This rapid stretching smoothed out the cosmos and planted the seeds for all the galaxies we see today. To understand how this happened, physicists study the "inflaton," a hypothetical field that drove this expansion. One of the most successful theories for this field is the Starobinsky model, which predicts specific patterns in the leftover radiation from the Big Bang. However, recent, incredibly precise measurements from the Atacama Cosmology Telescope have suggested a slight mismatch with these predictions. The data hints that the universe might have expanded slightly differently than the simplest version of the model suggests, creating a small but significant gap between theory and observation.

To bridge this gap, scientists have proposed that invisible quantum effects, known as radiative corrections, might be tweaking the behavior of the inflaton field. These are tiny adjustments that arise because the field interacts with other particles and forces, much like a boat moving through water experiences drag that isn't present in a vacuum. The question facing the physics community is whether these tiny quantum nudges are strong enough to fix the mismatch with the new telescope data. A team of researchers from Thailand has now investigated this possibility by looking at a specific, sophisticated version of gravity called weakly nonlocal gravity. Their work reveals that the size of the correction is not the most important factor; rather, it is the mathematical nature of the correction that determines whether it can actually change the outcome.

The researchers developed a new test to determine if a proposed quantum correction is "removable." In the language of their theory, a correction is removable if it can be absorbed by simply rescaling the height of the energy landscape or shifting the starting position of the inflaton field. If a correction is removable, it does not actually change the observable predictions of the universe; it is like changing the units of measurement on a ruler without changing the length of the object being measured. The team found that the quantum corrections arising purely from the gravitational sector in their model are indeed removable. This means that the gravitational running, or the way gravity changes at different energy scales, passes through the Starobinsky model without altering its successful predictions. The model remains robust against these specific quantum effects, preserving its original agreement with older data but failing to move toward the new, preferred values suggested by the latest telescope observations.

The study goes further to show that this result is not just a lucky accident but a structural feature of the theory. Because the corrections are removable, they evolve the model along a path where the predictions for the universe's expansion remain unchanged. The researchers calculated that even if they pushed the parameters to their absolute limits, the maximum shift these gravitational corrections could produce is far too small to explain the discrepancy seen in the new data. The effect is roughly twenty-five times smaller than what is needed to fix the mismatch. This leads to a firm conclusion: the tension between the Starobinsky model and the new telescope data cannot be resolved by the quantum behavior of gravity alone within this specific framework.

If the discrepancy is real, the researchers argue that the solution must lie elsewhere. They examined other proposed fixes, such as adding new terms to the equations of gravity or including the effects of matter particles like the Higgs boson. Their analysis shows that corrections involving matter are generally not removable; they cannot be hidden by a simple shift or rescaling. Consequently, these matter-induced effects can and do shift the predictions, but the team found that even these shifts are often too small or of the wrong sign to solve the problem without breaking other parts of the theory. The only way to get the required shift from matter would be to introduce couplings so large that they would likely destroy the mathematical consistency of the model.

The paper also addresses a common alternative approach where scientists choose a different scale to measure the quantum effects. The researchers demonstrated that choosing a scale based on the curvature of space, rather than the energy of the field, forces the corrections to be removable and thus harmless. Choosing a different scale might make the corrections look larger, but it introduces unphysical artifacts that make the theory unreliable. By sticking to the physically motivated scale, the team confirmed that the gravitational sector is effectively silent regarding this specific problem. The only way the model could be saved is if the discrepancy stems from a fundamental change in the theory's structure at the most basic level, or from contributions that are not purely gravitational.

Ultimately, this work provides a clear diagnostic tool for the future of inflationary theory. It establishes that for a wide class of advanced gravity theories, the quantum corrections generated by gravity itself are too weak and too well-behaved to explain the new observational data. The researchers have shown that the "removability" of a correction is a decisive property: if a correction can be removed by a simple mathematical shift, it cannot save a model from conflicting with data. Since the gravitational corrections in this class of theories are removable, they cannot be the solution. The burden of explaining the new data now falls on more radical possibilities, such as new particles, different initial conditions, or a completely different structure of gravity that does not possess this protective symmetry. The study leaves the door open for these other avenues while firmly closing the window on the idea that standard quantum gravitational running can fix the Starobinsky model.

Technical Summary: Removability Criterion for Radiative Corrections to the Starobinsky Attractor in Weakly Nonlocal Gravity

Problem Statement
The Starobinsky model ( $R + R^2$ ), a leading candidate for cosmic inflation, predicts scalar spectral index $n_s \simeq 1 - 2/N$ and tensor-to-scalar ratio $r \simeq 12/N^2$ . Recent observational data from ACT DR6, combined with Planck lensing and DESI, favor a value of $n_s \simeq 0.974$ , which is approximately $2\sigma$ higher than the standard Starobinsky prediction at $N_* \simeq 55$ . While previous studies suggested that percent-level radiative corrections could shift $\alpha$ -attractor and Starobinsky-like models into the favored region, the specific mechanism and stability of such shifts within a UV-complete framework remained unclear. This paper investigates whether radiative corrections in weakly nonlocal, super-renormalizable quantum gravity can resolve this tension, or if the model's predictions are protected against such shifts.

Methodology
The authors employ the slow-roll $N$ -formalism, where cosmological observables ( $n_s, r$ ) are functionals of the first slow-roll parameter $\epsilon(N)$ alone. The analysis proceeds through the following steps:

Framework: The study utilizes weakly nonlocal gravity completions of the Starobinsky model (following Krasnikov, Kuz'min, Tomboulis, and Modesto). In this framework, the action contains entire functions of the d'Alembertian $\Box$ devoid of zeros, ensuring super-renormalizability and preserving the classical Starobinsky background exactly.
Renormalization Group (RG) Flow: The authors calculate the one-loop running of local couplings ( $M_P, M_S, \xi, \alpha_C, \alpha_R, \rho_\Lambda$ ). Crucially, they adopt a physical improvement scale $\mu = H(\phi)$ , where $H$ is the Hubble parameter. This choice establishes an exact identity relating the RG time $t$ to the logarithm of the tree-level potential: $t = \frac{1}{2} \ln(V_{St}(\phi)/V_{St}(\phi_0))$ .
The Removability Criterion: The core theoretical tool is a differential criterion derived from the $N$ -formalism. The authors demonstrate that if a deformation $\Delta$ to the potential satisfies $(V/V')\Delta' = \text{const}$ , the resulting shift in observables is "removable." Such deformations can be absorbed by a rescaling of the potential amplitude and a constant shift of the inflaton field, leaving the leading-order predictions for $n_s$ and $r$ invariant.
Application: The criterion is applied to:
- The gravitational sector's one-loop flow in weakly nonlocal gravity.
- Matter-induced corrections (Coleman-Weinberg type).
- Various deformations proposed in the literature to address the $n_s$ tension (e.g., $R^n$ terms, non-minimal couplings).

Key Contributions and Results

The Removability Criterion: The paper establishes that the decisive property of a radiative correction is not its magnitude but its "removability." Deformations satisfying $(V/V')\Delta' = \text{const}$ preserve the attractor predictions. The authors derive this criterion explicitly, showing it corresponds to the tangent space of the E-model family $V \propto (1 - e^{-\sqrt{2/3}\phi/M_P})^p$ at $p=2$ .
Gravitational Sector Stability: In weakly nonlocal completions, the divergent local beta functions are one-loop exact. The RG flow, improved at $\mu = H$ $μ = H$ , evolves the model along the E-model family with an exponent $p = 2 + K$ $p = 2 + K$ , where $K$ $K$ is a combination of running couplings. Because the E-model predictions are exponent-independent at leading order, the gravitational running is removable.
- The residual shift in $n_s$ arises only from subleading terms and scales as $\mathcal{O}(\ln N / N^2)$ .
- Numerical estimates show the maximum induced shift is $|\Delta n_s| \lesssim 3.7 \times 10^{-4}$ , which is a factor of 25 too small to resolve the tension with ACT DR6.
Matter-Induced Corrections: In contrast to the gravitational sector, matter-induced corrections (specifically those involving field-dependent masses or non-minimal couplings) generally fail the removability test. They generate logarithms of the field rather than the potential, leading to non-removable shifts of order $\mathcal{O}(1/\ln N)$ or larger. This explains why previous literature found that matter couplings could shift predictions significantly.
Standard Model Constraints: For the Standard Model, the contribution to the removable channel is fixed by the Higgs non-minimal coupling $\xi_H$ . Due to the large value of the $R^2$ coefficient fixed by the measured scalar amplitude ( $\omega_{0,R} \approx 5.1 \times 10^8$ ), the matter contribution to the removable channel is negligible ( $\Delta n_s \sim 10^{-14}$ ).
Evaluation of Proposed Deformations: The criterion is applied to deformations proposed since ACT DR6:
- $R^n$ terms in $f(R)$ : Only $n=2$ is removable. Higher powers ( $n > 2$ ) are non-removable and shift $n_s$ as $\mathcal{O}(N^{n-1})$ .
- Marginal deformations and field logarithms: These fail the test and shift $n_s$ significantly.
- The paper demonstrates that no proposed deformation capable of raising $n_s$ to the ACT preferred value belongs to the removable class.

Significance and Claims
The paper claims that the tension between the Starobinsky model and ACT DR6 data cannot be resolved by local one-loop radiative corrections within the class of weakly nonlocal, super-renormalizable quantum gravity.

Structural Protection: The gravitational sector is structurally protected by the removability criterion. The specific choice of improvement scale $\mu = H$ ensures that the RG flow tracks the potential, rendering gravitational corrections invisible at leading order.
Source of Shifts: If a significant shift in $n_s$ is observed, it must originate from non-removable sources: matter-induced corrections, post-inflationary history (reheating), or tree-level structural modifications outside the weakly nonlocal class.
Predictive Power: The criterion provides a model-independent test for any proposed deformation. It distinguishes between corrections that merely rescale the model (and thus preserve predictions) and those that fundamentally alter the attractor behavior.
Tensor Sector: While the scalar sector is protected, the tensor sector ( $r$ ) is not protected by the same mechanism, though the running effects on the tensor tilt are negligible ( $\mathcal{O}(10^{-12})$ ). The amplitude of $r$ remains governed by tree-level form-factor parameters.

In conclusion, the authors assert that within the specified theoretical framework, the Starobinsky attractor is robust against the specific type of radiative corrections generated by the gravitational sector. The observed discrepancy with ACT data, if confirmed, points toward physics beyond the local one-loop running of this specific class of completions.

Technical Summary: Removability Criterion for Radiative Corrections to the Starobinsky Attractor in Weakly Nonlocal Gravity

More like this