Hubble Tension as an Effect of Horizon Entanglement Nonequilibrium

TL;DR

The paper introduces HEED, an infrared horizon-entanglement deficit mechanism that generates a smooth energy density with $ρ_{HEED} ∝ H^2/G$, activated at late times to affect the expansion history without perturbing recombination. It formalizes the deficit via $δ(a)=1-S_{ent}(a)/S_{BH}(a)$ and derives the background dynamics, an instantaneous equation of state, and the linear growth for a non-clustering entanglement sector through a minimal three-parameter activation $\{c_{e0}^2, a_t, k\}$. A Bayesian analysis using SN, BAO, CC, and RSD data shows nonzero $c_{e0}^2$ with activation around $a_t ∼ 0.4$–$0.6$, producing an effective late-time Hubble rate $H_0^{eff} ≈ 73$ km s$^{-1}$ Mpc$^{-1}$ and a mild suppression of $fσ_8$, while keeping distance measures broadly consistent with $Λ$CDM. Although current data do not decisively distinguish HEED from $Λ$CDM, the framework provides a consistent, locally anchored interpretation of the Hubble tension and makes testable predictions for ISW and growth that future surveys can probe. The work highlights a quasi-local horizon-based IR modification as an alternative to early-universe changes and outlines directions to connect $δ(a)$ to microphysical horizon information concepts.

Abstract

We propose an infrared mechanism for alleviating the Hubble constant tension, based on a small departure from entanglement equilibrium at the cosmological apparent horizon. If the horizon entanglement entropy falls slightly below the Bekenstein-Hawking value, we parametrize the shortfall by a fractional deficit $δ(a)$ evolving with the FLRW scale factor $a$. The associated equipartition deficit at the Gibbons-Hawking temperature then sources a smooth, homogeneous component whose density scales as $H^{2}/G$, with a dimensionless coefficient $c_{e}^{2}(a)$ of order unity times $δ(a)$. Because this component tracks $H^{2}$, it is negligible at early times but can activate at redshifts $z\lesssim 1$, raising the late time expansion rate by a few percent without affecting recombination or the sound horizon. We present a minimal three parameter activation model for $c_{e}^{2}(a)$ and derive its impact on the background expansion, effective equation of state, and linear growth for a smooth entanglement sector. The framework predicts a small boost in $H(z)$, a mild suppression of $fσ_{8}(z)$, and a corresponding modification of the low-$z$ distance-redshift relation. We test these predictions against current low-redshift data sets, including SN~Ia distance moduli, baryon acoustic oscillation distance measurements, cosmic chronometer $H(z)$ data, and redshift space distortion constraints, and discuss whether the $H_0$ tension can be consistently interpreted as a late-time, horizon-scale information deficit rather than an early universe modification.

Figures (8)

• Figure 1: HEED's linear growth impact illustrated for three sets of the activation parameters.
• Figure 2: HEED corrected low redshift Hubble constant related via its residuals relative to the $\Lambda$CDM calculations.
• Figure 3: Triangle plot for the HEED parameterization showing the marginalized posterior distributions of $\{H_0,\Omega_m,c_{e0}^2,a_t,k,\sigma_8\}$ obtained from the joint SN+BAO+CC+RSD likelihood. Shaded contours indicate the 68% and 95% credible regions, while the diagonal panels show the corresponding one--dimensional marginals. In the anchored analysis, the external Gaussian prior is applied to the effective late--time Hubble scale $H_0^{\rm eff}=H_0/\sqrt{1-c_{e0}^2}$ rather than directly to $H_0$, which induces the characteristic degeneracy direction in the $(H_0,c_{e0}^2)$ plane.
• Figure 4: Pantheon+ SN Ia distance moduli $\mu(z)$ compared to the best--fit HEED (solid) and $\Lambda$CDM (dashed) predictions. The fits are shown after profiling over the nuisance offset $\Delta M$ (absolute-magnitude/zero-point) as described in the text.
• Figure 5: CC measurements of $H(z)$ with $1\sigma$ uncertainties compared to the best--fit HEED (solid) and $\Lambda$CDM (dashed) predictions.