A generic omega b tension in early-time solutions to the Hubble tension

arXiv:2604.05095 · astro-ph.CO, hep-ph · Submitted 2026-04-06 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.

Vera: Next we'll be talking about the paper "A generic omega b tension in early-time solutions to the Hubble tension".

Jocelyn: The paper was written by Cara Giovanetti from Lawrence Berkeley National Laboratory and University of California, Berkeley Institute for Theoretical Physics (Leinweber Institute).

Vera: Stay tuned as we take you through the paper and discuss its implications.

Introduction to A Generic omega b Tension in Early-time Solutions to the Hubble Tension: Vera: We are looking at a very interesting paper titled "A generic omega b tension in early-time solutions to the Hubble tension," and it addresses one of the biggest problems in modern cosmology: how do we reconcile the local measurement of H zero with our observations of the early universe?

Jocelyn: The authors are essentially proposing that if we want to boost H zero using models where physics happened very early on, there' a catch, and this paper aims to uncover that catch.

Subrahmanyan: It’s not just a simple measurement error they suggest; the theoretical framework of these models is fundamentally linked to the density parameters in a way that creates unavoidable constraints.

Vera: The premise of any early-time solution is that some new physics—like extra radiation or dark energy—is present before recombination, but this paper shows that this mechanism doesn' omega b (the baryon density) is inherently tied to the scaling of H zero.

Jocelyn: If you are trying to increase the expansion rate, the paper suggests you are simultaneously forcing a higher concentration of baryons than what we know exists.

Subrahmanyan: This relationship is predicted by how angular scales in the CMB, like A and eq, respond to changes in H(z), meaning they cannot be adjusted independently of the baryon density.

Vera: So, when we discuss "A generic omega b tension," we're talking about a systematic failure for any model that tries to solve the Hubble discrepancy through early modifications.

Jocelyn: It’s a clear signal that any attempt to fix H zero via these early-universe models introduces a fundamental conflict with our understanding the baryon density.

Subrahmanyan: This opens up a huge discussion about where the limits of current cosmological theory lie when we push for higher values of the Hubble constant.

Vera: We need to look at how this tension manifests in specific data sets before we move on to what the authors found in their detailed analyses.

Summary of Findings from A Generic omega b Tension in Early-time Solutions to the Hubble Tension: Jocelyn: Now, let’s look at the core findings summarized by Giovanetti et al. in "A generic omega b tension in early-time solutions to the Hubble tension," and see exactly what that conflict looks like when it hits real data.

Vera: The paper shows that these early-time models generically prefer a much higher value of omega b, which is the baryon density, than we've ever seen before.

Subrahmanyan: This is significant because this preference isn't just a random fluctuation; it’s a robust consequence of how the angular diameter distance at decoupling, D A(z*), scales with the Hubble parameter H(z).

Jocelyn: We know from Big Bang Nucleosynthesis, or BBN, that omega b is extremely tightly constrained by the abundance of primordial deuterium.

Vera: The paper demonstrates that these high- omega b predictions are in direct conflict with this precise BBN constraint at the percent level.

Subrahmanyan: It's a systematic issue where the initial conditions required to achieve a large H zero inevitably lead to a baryon density that is inconsistent with how elements formed in the first few minutes.

Jocelyn: The analysis shows that this conflict isn't just theoretical; it’s seen when combining CMB data with BBN likelihood components, which dramatically lowers the preferred value of H zero.

Vera: This suggests that a simple "boost" to the early-time physics won't work if we respect our knowledge of primordial nucleosynthesis.

Subrahmanyan: The constraints are simply too tight; the universe seems to have only had a certain amount of matter available when the first elements were created.

Jocelyn: It’s a powerful demonstration that any simple fix for H zero will clash with our understanding of the first moments of cosmic nucleosynthesis.

Vera: But this problem isn' clearly solved by simply adding BBN to what we already have in the data, so how do we even begin to fix this?

Proposed Improvements in A Generic omega b Tension in Early-time Solutions to the Hubble Tension: Vera: The authors then pivot and look at how these models perform when they are analyzed with a full BBN likelihood, which is detailed in "A generic omega b tension in early-time solutions to the Hubble tension."

Jocelyn: They examine two specific models, WZDR and EDE, and show that adding the BBN constraint drastically changes the results compared to analyses that ignore it.

Subrahmanyan: The inclusion of BBN acts as a strong brake on any attempt to achieve a high H zero using these early-time solutions because of the tight coupling between omega b and H zero.

Vera: For example, in their analysis of WZDR, when they include the BBN likelihood, the maximum value for H zero drops substantially compared to when they exclude it.

Jocelyn: It’s fascinating because even though these models are designed to solve the Hubble tension, they can't achieve that high H zero once you factor in our knowledge of deuterium.

Subrahmanyan: This shows that the parameters are highly interdependent; you can't just change one input and expect the output to remain consistent with other established physics.

Vera: The work in "A generic omega b tension in early-time solutions to the Hubble tension" highlights that reconciling these models with BBN requires finding a way around this specific coupling between omega b and H zero.

Jocelyn: This definitely points toward needing some very clever interventions or perhaps looking at more complex model combinations to achieve consistency.

Subrahmanyan: The authors' approach is not to discard the models, but to show that achieving concordance requires a highly detailed understanding of how these new physics interact with the constraints imposed by early cosmic chemistry.

Vera: We need a path forward that respects both the large-scale structure we see and what happened in those first critical moments of matter.

Conclusion on A Generic omega b Tension in Early-time Solutions to the Hubble Tension: Vera: So, we have covered a lot of ground with "A generic omega b tension in early-time solutions to the Hubble tension," and the main message is that any attempts to solve H zero using early dark energy or similar mechanisms face a fundamental conflict with BBN.

Jocelyn: That’s a huge challenge for our survey planning, as it means we cannot simply assume that we can reach higher values of the Hubble constant while maintaining consistency with our precise measurements of primordial deuterium.

Subrahmanyan: I think the implications are massive because this is more than just a minor issue; it strongly suggests that solving the Hubble tension will require a deep theoretical shift in how we model early-universe physics, perhaps involving much more intricate couplings.

Vera: I agree, Subrahmanyan, and the fact that this conflict is so "generic" makes it a very important finding—it’s not just an accidental failure of one specific model.

Jocelyn: It really forces us to think about how our combined analyses need to be even more sensitive when we are trying to constrain both H zero and omega b simultaneously in future experiments.

Subrahmanyan: I'm excited to see if the next generation of observations can really put pressure on whether or these modified cosmologies are the true picture.

Vera: We’ve covered a lot of ground today, and this tension is certainly going to keep us thinking about how we approach the Hubble constant problem.

Jocelyn: It really makes us wonder what other physical mechanisms could provide a cleaner, more consistent path toward concordance between BBN and late-time measurements.

Lawrence Berkeley National Laboratory · University of California, Berkeley Institute for Theoretical Physics (Leinweber Institute)

astro-ph.CO, hep-ph

Submitted: 2026-04-06

Updated: 2026-09-03

Journal ref: JCAP 08 (2026) 074

DOI: 10.1088/1475-7516/2026/08/074

Code: https://github.com/CobayaSampler/planck_supp_data_and_covmats

License: http://creativecommons.org/licenses/by-sa/4.0/

Importance score: 35/100

The gist: The paper investigates a potential tension in the baryon density parameter (omega b) that arises when applying early-time physics solutions intended to resolve discrepancies in measurements of the

Key concepts

Hubble Tension
The core problem of modern cosmology, this is the conflict between local measurements of the universe's expansion rate and observations made in the early universe.
ωb (Baryon Density)
This parameter measures the concentration of ordinary matter in the universe. The paper shows that attempts to solve Hubble tension by increasing H_zero inevitably force a higher value for this density.
Big Bang Nucleosynthesis (BBN)
The process where elements formed in the first few minutes of existence. BBN provides a very tight constraint on ωb based on the observed abundance of primordial deuterium.

Terminology

Summary

The paper investigates a potential tension in the baryon density parameter (omega b) that arises when applying early-time physics solutions intended to resolve discrepancies in measurements of the Hubble constant (H 0). By analyzing how various modified cosmological models—such as WZDR (Wiggle-Zeta Dark Radiation) and EDE (Early Dark Energy)—impact standard constraints derived from Planck, BAO, and distance ladder measurements, the authors aim to determine if this tension is generic across different early-universe scenarios.

The Standard Cosmological Baseline

The analysis establishes a robust baseline using established data combinations. The primary likelihoods employed include:

  • Planck 2018 (TT, TE, EE+lowE and lensing) [1].

  • BAO measurements from BOSS DR12 [13] and 6dF [41].

  • Cosmic distance ladder measurements from Pantheon [16] and SH0ES [2].

  • The BBN likelihood, which is described in the text.

These combined data sets yield best-fit values for key parameters, such as ch squared and omega b. The inclusion of the BBN likelihood significantly constrains the parameter space, influencing the derived values for H 0 and other fundamental constants.

Impact of Early Dark Energy (EDE) Models

The EDE model introduces specific parameters unique to its framework, which are critical for understanding deviations from standard CDM. These parameters include:

  • f EDE: Controlling the fraction of dark energy added.

  • 10 a c: Controlling the scale factor at which dark energy dominates.

  • i: Determining the initial displacement of the scalar field.

The analysis shows that when considering EDE, constraints on omega b and H 0 are affected. For instance, the inclusion of BBN likelihoods causes shifts in these parameters compared to analyses without BBN constraints, suggesting a strong interplay between early-time physics and nucleosynthesis predictions.

Effect of WZDR (Wiggle-Zeta Dark Radiation)

The WZDR model introduces an extra radiation component at late times, quantified by the parameter NIR. This modification alters the cosmological fit and is tested in conjunction with BBN constraints. The results demonstrate how this late-time energy injection affects the derived values for fundamental parameters. The comparison between WZDR analyses with and without BBN likelihoods highlights that incorporating nucleosynthesis data imposes stringent limits on the allowed parameter space, particularly concerning omega b.

Baryon Density Tension (omega b)

The central focus of the paper is the potential generic omega b tension. The derived best-fit values for omega b are compared across various models. The data illustrate that:

  • The incorporation of BBN likelihoods consistently influences the preferred value for omega b.

  • The shift in parameters like H 0 and omega b suggests that early-time physics solutions designed to resolve the Hubble tension may simultaneously introduce or exacerbate tensions in other fundamental parameters, such as the baryon density.

In summary, the paper provides a comprehensive parameter constraint analysis across multiple modified cosmological models. The results emphasize that any proposed solution to the H 0 tension must be rigorously tested against independent constraints, particularly those derived from Big Bang Nucleosynthesis (BBN), which appear to constrain omega b in a manner that is model-dependent.

Improvements for AI systems

This document details improvements for developing specialized AI systems tailored for high-dimensional, complex parameter inference in observational cosmology. The goal is to enhance the speed, reliability, and depth of model comparison beyond current state-of-the-art MCMC techniques.

The core improvement is moving from sequential statistical sampling (like MCMC) to a hybrid architecture that combines advanced surrogate modeling with deep generative inference networks, allowing for real-time exploration of vast likelihood spaces and automated detection of model degeneracy.

  • Current Limitation: Model comparison relies on metrics like AIC or chi 2 tot, which are computationally expensive to calculate accurately across multiple, non-nested models (e.g., comparing the standard CDM model to WZDR or EDE).

  • AI Improvement: Implement a Variational Autoencoder (VAE) trained on the posterior distributions of candidate cosmological parameters. The VAE learns a compressed, low-dimensional latent space representation (z) that captures the essential variance of each model's likelihood function.

  • What the Improved System Can Do:

  • Automated Model Ranking: Instead of calculating AIC or Bayes Factors through extensive sampling, the BMCS can rapidly estimate the relative evidence between two competing models (M A vs M B) by comparing their respective latent space manifolds (z A and z B).

  • Degeneracy Mapping: It will automatically detect subtle, non-linear parameter degeneracies (e.g., the interplay between tau reio and A s) that might be missed by simple 2D contour plots, projecting these degeneracies onto the latent space for clearer visualization and constraint identification.

  • Current Limitation: Calculating the likelihood function L(D theta) requires running complex simulations (e.g., calculating chi squared based on Planck TT, TE, EE data across multiple multipoles). This is the primary computational bottleneck.

  • AI Improvement: Utilize a Gaussian Process (GP) Regression framework coupled with Physics-Informed Neural Networks (PINNs). The PINN component is crucial: it enforces known physical relationships (e.g., the relationship between c h squared and the CMB acoustic peak positions) as hard constraints within the loss function, guiding the GP interpolation.

  • What the Improved System Can Do:

  • Instantaneous Likelihood Evaluation: Instead of running full Boltzmann solvers (like CAMB/CLASS) for every point in parameter space, the HDLSM builds a surrogate model that predicts L(D theta) almost instantly from theta. This reduces the inference time from days/hours to minutes.

  • Accelerated Sampling: The GP provides a highly accurate, differentiable approximation of the likelihood landscape, allowing standard sampling algorithms (e.g., Hamiltonian Monte Carlo) to proceed with vastly increased efficiency and convergence speed, reducing systematic errors associated with insufficient sampling.

  • Current Limitation: The analysis relies on combining disparate datasets (CMB, BAO, Supernovae). Discrepancies (like the H 0 tension or tension between early/late universe measurements) are often presented as separate results.

  • AI Improvement: Implement a Graph Neural Network (GNN) structure. Each dataset (Planck TT, Pantheon, BAO) and each parameter (c h squared, H 0) is treated as a node in the graph, and the known physical relationships or statistical correlations between them are treated as weighted edges.

  • What the Improved System Can Do:

  • Systematic Tension Quantification: The GNN analyzes the consistency of constraints across all nodes simultaneously. If a conflict exists (e.g., H 0 from local measurements disagrees with H 0 derived from CMB), the GNN doesn't just report the discrepancy; it quantifies the degree of statistical incompatibility (Incompatibility Score) and proposes plausible missing link nodes (e.g., a new physics parameter like f EDE) that, if added, minimize this overall graph tension score.

  • Outlier Flagging: It can automatically flag parameters or datasets whose derived constraints deviate significantly from the established network consensus, providing immediate alerts for potential systematic errors in the underlying observational data or analysis pipelines.

Abstract

I show that early-time (pre-recombination) solutions to the Hubble tension are generically expected to increase the preferred baryon density ω b. This puts these models in tension with Big Bang Nucleosynthesis (BBN), as measurements of primordial deuterium constrain ω b at percent level. I show that existing analyses are in tension with the BBN determination of ω b, and that including a likelihood component for primordial deuterium deters two representative models from recovering a high H 0.

Sources

Related papers