Observational constraints on the modified cosmology inspired by string T-duality

TL;DR

This paper tests a string T-duality–inspired cosmology in which a zero-point length $l_0$ induces a corrected thermodynamic relation at the apparent horizon, yielding a modified Friedmann equation with a dimensionless coupling $\beta \sim l_0^2 H_0^2$. Using a Cobaya-based Bayesian analysis with six data combinations of PantheonPlus, Union3, cosmic chronometers, DESI DR2 BAO, and Amati-calibrated GRBs, the authors constrain $\beta$ and compare to $\Lambda$CDM via AIC. They find $\beta$ is tightly bounded, $\beta \lesssim \mathcal{O}(10^{-3})$ when BAO data are included, and the $\Lambda$CDM and T-duality models yield statistically equivalent fits (AIC differences around zero to a few units). The results demonstrate that current late-time cosmological data place strong limits on quantum-gravity–induced corrections in the late Universe, while BAO data provide the strongest constraint, and GRBs offer limited additional information. The work sets the stage for future joint early- and late-time analyses (e.g., CMB) to more stringently test minimal-length cosmologies and motivates comparative studies with alternative minimal-length frameworks such as the $q$-metric or GUP formalisms.

Abstract

We explore the cosmological consequences of a modified cosmology inspired by string T-duality. We incorporate the zero-point length correction, $l_0$, into the gravitational potential and derive the modified Friedmann equations via thermodynamic approach at the apparent horizon of a Friedmann-Robertson-Walker (FRW) universe. The resulting framework introduces a dimensionless coupling parameter $β\sim l_0^2H_0^2$ quantifying deviations from the standard $Λ$CDM model. Using Bayesian inference with \textsc{Cobaya} and MCMC sampling, we constrain the model parameter against late-time observations, including PantheonPlus and Union3 Type~Ia supernovae, cosmic chronometers, DESI~DR2 BAO measurements, and Amati-calibrated GRBs. The joint analysis yields an upper bound $β\lesssim \mathcal{O}(10^{-3})$ (68\% C.L.), implying that departures from $Λ$CDM are extremely small within current precision. Model comparison through the Akaike Information Criterion shows that the $Λ$CDM and T-duality models provide statistically equivalent fits to the data, exhibiting only a marginal preference for $Λ$CDM. These results provide the first quantitative observational constraints on string T-duality inspired modified cosmology and underscore the potential of future high-precision surveys to test quantum-gravity induced corrections in a late-time universe.

Figures (2)

• Figure 1: Confidence regions for the free parameters of the string T-duality cosmology, derived from the joint likelihood of the combined datasets including the PP SNIa catalogue.
• Figure 2: Confidence regions for the free parameters of the string T-duality cosmology, derived from the joint likelihood of the combined datasets including the U3 SNIa catalogue.