⚛️ general relativity

A Case for an Inhomogeneous Einstein-de Sitter Universe

This paper proposes an inhomogeneous Einstein-de Sitter universe model where cosmic acceleration arises from structure formation rather than dark energy, demonstrating that this framework fits observational data as well as $\Lambda$ CDM while resolving the $H_0$ tension and yielding a consistent cosmic age.

Original authors: Peter Raffai, Dominika E. R. Kis, Dávid A. Ködmön, Adrienn Pataki, Rebeka L. Böttger, Gergely Dálya

Published 2026-01-29

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Peter Raffai, Dominika E. R. Kis, Dávid A. Ködmön, Adrienn Pataki, Rebeka L. Böttger, Gergely Dálya

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

Imagine the universe as a giant, expanding balloon. For decades, the standard story (called the ΛCDM model) has been that this balloon is being blown up by an invisible, mysterious force called Dark Energy. Without this force, the balloon should be slowing down because gravity is pulling everything together. But when we look at the stars, we see the balloon is actually speeding up. So, scientists invented Dark Energy to explain the push.

This new paper, by Peter Raffai and his team, tells a different story. They propose that we don't need a mysterious "push" at all. Instead, they suggest the universe is speeding up because of clumps and bumps in the fabric of space itself.

Here is the breakdown of their idea, using simple analogies:

1. The "Smooth vs. Bumpy" Universe

The Old View (Smooth): The standard model assumes the universe is like a perfectly smooth, flat sheet of rubber. It expands evenly everywhere, like a balloon inflating in a vacuum.
The New View (Bumpy): The authors say the universe is more like a crumpled piece of paper or a rough, bumpy road. It has valleys (dense clusters of galaxies) and hills (empty voids).

2. The "Traffic Jam" Analogy

Imagine a highway where cars (galaxies) are driving.

In the smooth model: Everyone drives at the exact same speed, and the road stretches out evenly.
In the bumpy model: Some cars get stuck in traffic jams (dense areas), while others zoom through empty stretches (voids).

The authors argue that because the universe is "bumpy," the empty spaces expand faster than the crowded spaces. As the universe gets older, these empty spaces grow larger and take up more of the total volume. It's like if you had a bag of marbles and sand; as the sand spreads out, it eventually takes up most of the bag, pushing the marbles aside.

3. The "Illusion" of Acceleration

Here is the magic trick: Because the empty, fast-expanding parts of the universe are getting bigger and bigger, they start to dominate the view. When we look out from Earth, we are essentially looking through these expanding "empty corridors."

The authors claim that this structural growth creates an illusion of acceleration. It's not that a mysterious force is pushing the universe; it's that the "fast lanes" of the universe are becoming the main road. The math shows that this natural clumping can mimic the effects of Dark Energy perfectly, without needing to invent a new substance.

4. Solving the "Speedometer" Problem

There is a famous disagreement in astronomy called the Hubble Tension.

Method A (The Baby Picture): Looking at the Cosmic Microwave Background (the "baby picture" of the universe) suggests the universe is expanding at a slower speed.
Method B (The Adult Picture): Looking at nearby supernovas (exploding stars) suggests it's expanding faster.

The standard model struggles to make these two numbers match. The authors tested their "Bumpy Universe" model against the data. They found that their model fits the "baby picture" data just as well as the standard model, but it naturally matches the "adult picture" speed without any fudging. It resolves the conflict by suggesting the universe is expanding at a speed that sits comfortably between the two conflicting measurements.

5. The "Coasting" Future

In the standard model, the universe is accelerating toward a cold, dark end where everything is ripped apart or frozen.
In this new model, the universe is "coasting." Imagine a car that has been pushed and is now rolling down a long, gentle hill. It's not speeding up wildly, and it's not stopping; it's just rolling at a steady, linear pace. The authors call this a "Milne state." It suggests a future where the universe keeps expanding forever, but in a calm, steady rhythm, rather than a chaotic explosion.

The Bottom Line

The paper argues that Dark Energy might not exist. Instead, the acceleration we see is a side effect of the universe becoming more lumpy and structured over time.

Does it fit the data? Yes. They tested it against the most precise maps of the universe (CMB), galaxy surveys (BAO), and exploding stars (Supernovae). It fits almost as well as the standard model.
Does it solve the age problem? Yes. The standard model sometimes predicts a universe that is too young to hold the oldest stars we see. This new model predicts an age (13.67 billion years) that matches the age of the oldest star clusters perfectly.

In short: The universe isn't being pushed by a ghost; it's just getting bumpy, and those bumps are making it look like it's speeding up.

Technical Summary: A Case for an Inhomogeneous Einstein-de Sitter Universe

Problem Statement
The prevailing $\Lambda$ CDM cosmological model, while successful across multiple epochs, faces significant tensions between locally measured parameters and those derived from early-universe observations (CMB and BAO). Most notably, there is a discrepancy in the Hubble constant ( $H_0$ ) and the structure growth parameter ( $S_8$ ). The standard model relies on a global-to-local approach, assuming large-scale homogeneity and isotropy (the cosmological principle) and attributing cosmic acceleration to a dark energy component (a cosmological constant, $\Lambda$ ). This paper challenges the necessity of dark energy, proposing that observed acceleration may be an emergent phenomenon arising from local structure formation within an inhomogeneous universe, rather than a global property driven by a new energy component.

Methodology
The authors propose a "local-to-global" framework based on two principles:

The global evolution of the universe follows the Friedmann equations with a global scale factor $a(t)$ .
This global evolution is the volume-weighted ensemble average of finite-volume regions, each evolving according to its own Friedmann equations with local scale factors $a_i(t)$ and parameters $\{\theta_i\}$ .

Crucially, the authors reject the standard assumption that all probed regions share identical cosmological parameters ( $\{\theta_i\} = \{\theta\}$ ). Instead, they model an inhomogeneous Einstein-de Sitter (iEdS) universe. In this model:

The universe is globally spatially flat but composed of regions with varying local matter densities and spatial curvatures.
Regions with positive curvature ( $\Omega_{k,i} > 0$ ) expand faster and gain disproportionate weight in the volume average as structure forms.
This differential expansion generates an effective negative spatial curvature in the global dynamics, mimicking cosmic acceleration without a cosmological constant.
The model is parameterized phenomenologically using an effective curvature equation of state $w_K(a)$ and a dimensionless parameter $r^2$ representing the variance of local Hubble parameters ( $\sigma_H^2/H^2$ ).

The authors tested the iEdS model against three primary datasets:

Planck 2018 CMB: Temperature power spectrum.
DESI DR2 BAO: Baryon Acoustic Oscillation data.
Pantheon+ SNe Ia: Type Ia supernova distance moduli.

The fitting procedure involved fixing the Planck constraint $\Omega_{M,0}H_0^2 = 1431.354$ km $^2$ s $^{-2}$ Mpc $^{-2}$ and varying $H_0$ and the structural parameter $r^2$ to match observations.

Key Contributions

Theoretical Framework: The paper establishes a local-to-global derivation where cosmic acceleration arises purely from General Relativistic effects of structure formation (backreaction) in an inhomogeneous Einstein-de Sitter universe, eliminating the need for dark energy.
Quasilinear Coasting Evolution: The model predicts a transition from an early Einstein-de Sitter state to a final asymptotic Milne state ( $a(t) \propto t$ ), characterized by quasilinear coasting expansion driven by growing inhomogeneities.
Phenomenological Mapping: The authors derive analytic forms for the effective curvature density and equation of state, allowing the iEdS model to be tested against standard cosmological probes using existing Boltzmann codes (e.g., CAMB) by mapping it to an effective dark energy model.

Results

CMB Fit: The iEdS model fits the Planck 2018 CMB temperature power spectrum comparably to the flat $\Lambda$ CDM model. With $H_0 = 72.5$ km s $^{-1}$ Mpc $^{-1}$ and $\Omega_{M,0} = 0.272$ , the $\chi^2$ values are nearly identical (2568.6 for iEdS vs. 2570.2 for $\Lambda$ CDM), and residuals pass normality tests.
Hubble Tension Resolution: The iEdS model resolves the $H_0$ tension. It yields a best-fit $H_0 \approx 72.5$ km s $^{-1}$ Mpc $^{-1}$ , consistent with local measurements (SH0ES) and the CMB-derived value within the model's own framework, whereas the standard $\Lambda$ CDM fit to the same data yields a lower $H_0$ ( $\approx 67.3$ ), creating a $>4\sigma$ tension with local measurements.
BAO and SN Ia: The model fits DESI DR2 BAO and Pantheon+ SNe Ia data. While $\Lambda$ CDM shows a slightly better Bayes factor for BAO data, the iEdS model removes the $H_0$ tension seen in $\Lambda$ CDM fits to both datasets.
Cosmic Age: The iEdS model yields a cosmic age of $t_0 \simeq 13.67$ Gyr, which is consistent with globular-cluster age estimates ( $13.6 \pm 0.3$ Gyr). In contrast, the $\Lambda$ CDM model fitted to Pantheon+ SNe yields an age of $12.4 \pm 0.3$ Gyr, creating a tension with stellar age estimates.
Statistical Caveats: The authors note that both models exhibit systematic overfitting of the Pantheon+ sample (standardized residual dispersion $\sigma = 0.95$ ), suggesting limitations in the data's uncertainty estimates or model assumptions rather than a definitive failure of the iEdS physics.

Significance and Claims
The paper claims that the observed cosmic acceleration can be explained as a purely general-relativistic effect of structure formation in an inhomogeneous universe, removing the need for dark energy. The iEdS model demonstrates that a universe starting with negligible inhomogeneities can naturally evolve toward an effective Milne state, resolving the $H_0$ tension and providing a cosmic age consistent with stellar populations.

The authors emphasize that this is a phenomenological realization of a broader local-to-global framework. They state that while the current model fits existing data comparably to $\Lambda$ CDM, the ultimate validation requires a realistic, first-principles theory of structure formation (specifically deriving the evolution of $\sigma_H^2/H^2$ ) through detailed relativistic simulations. If confirmed, such a model would imply a universe that asymptotically approaches linear expansion ( $a(t) \propto t$ ), resulting in infinite comoving horizons and a cosmos where every region is, in principle, observable, contrasting with the finite observable horizons of eternally accelerating $\Lambda$ CDM universes.