⚛️ general relativity

Kinetic-theory derivation and Bayesian analysis of the finite-temperature shifted CMB power spectrum

This paper employs Bayesian MCMC analysis and a custom kinetic-theory CMB pipeline to demonstrate that finite-temperature quantum-gravitational corrections to the cosmological constant, represented by $\Omega_{\Lambda_2}$ and $\Omega_{\Lambda_3}$ , remain undetected in Planck 2018 data because their subtle signatures are effectively absorbed by degeneracies with the Hubble constant and overall spectrum amplitude.

Original authors: I. Y. Park

Published 2026-10-06

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: I. Y. Park

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

The universe is not merely a static stage where galaxies drift apart; it is a dynamic system governed by a set of rules that dictate how space expands over time. For decades, the standard model of cosmology has relied on a simple idea to explain why this expansion is accelerating: a constant energy inherent to empty space itself, known as the cosmological constant. This concept has been remarkably successful, matching the detailed map of the cosmic microwave background—the faint afterglow of the Big Bang—with sub-percent precision. However, a persistent tension remains. Measurements of the universe's current expansion rate derived from this ancient light disagree with measurements taken from nearby galaxies. This discrepancy has led scientists to ask a fundamental question: is the vacuum energy truly constant, or does it evolve as the universe changes? One compelling possibility is that the vacuum energy responds to the temperature of the cosmos, much like how the pressure of a gas changes with heat. If this is true, the energy budget of the universe would contain subtle, temperature-dependent corrections that have so far been too small to detect.

A researcher set out to investigate this possibility by looking for two specific, tiny adjustments to the cosmological constant. These adjustments, predicted by quantum theory at finite temperatures, would act like extra components in the universe's energy mix. One would behave like radiation, fading away as the universe expands, while the other would behave like spatial curvature, fading more slowly. The researcher wanted to know if the precise measurements of the cosmic microwave background, taken by the Planck satellite, could reveal the fingerprints of these two new components. To do this, they did not rely on a single method. Instead, they combined two powerful approaches: a rigorous statistical analysis that accounts for all possible uncertainties, and a custom-built computer simulation that traces the physics of the early universe from first principles. The goal was to see if the data could distinguish these new effects from the standard model, or if they were simply hiding in the noise.

The researcher first built a new computational pipeline from scratch, following the kinetic theory of gases as applied to the early universe. Unlike standard software that switches between different mathematical approximations at various stages, this new tool integrated the full physics continuously, avoiding any potential errors introduced by switching methods. They used this tool to calculate exactly how the two new temperature-dependent terms would alter the pattern of temperature fluctuations in the cosmic microwave background. They found that the effect of the first term, which scales like radiation, is almost indistinguishable from a simple change in the overall brightness of the entire spectrum. It acts as a uniform dimming or brightening, which is already tightly constrained by other known factors. The second term, which scales like curvature, produces a more complex pattern of shifts in the peaks of the spectrum. However, the researcher discovered that this complex pattern is almost perfectly mimicked by a slight change in the Hubble constant, the number that describes how fast the universe is expanding today.

To test whether these effects could actually be detected, the researcher performed a Bayesian statistical analysis. This method explores the entire range of possible values for all cosmological parameters simultaneously, rather than just finding a single "best fit." They fed their modified model into the Planck 2018 data and let the computer sample millions of possibilities to see what the data truly allowed. The results were clear: the data showed no evidence for either of the two new parameters. Both were consistent with zero. The first parameter, the radiation-like term, was constrained to be smaller than about half a percent of the known radiation density. The second parameter, the curvature-like term, was only constrained from one side, meaning the data allowed it to be zero or slightly negative, but could not rule out larger positive values. The reason for this lack of detection was not a failure of the data, but a fundamental limitation of the physics. The signature of the second parameter was so closely tied to the Hubble constant that the two could trade places without changing the fit to the data. The part of the signal that could not be mimicked by the Hubble constant existed only at the largest scales of the universe, where the inherent randomness of the cosmos—known as cosmic variance—makes it impossible to measure with greater precision, regardless of how good the instruments are.

The study concludes that while the theoretical corrections to the cosmological constant are real and mathematically sound, they are currently undetectable with existing cosmic microwave background data. The researcher found that the apparent improvement in the fit seen in previous, simpler analyses was an illusion created by the fact that the data was being compared to a noise-free theoretical curve rather than the actual, noisy sky. When the full complexity of the data and the uncertainties of all other parameters are taken into account, the new parameters disappear into the background. The paper demonstrates that the two new terms are not simply a relabeling of known radiation or curvature; they are distinct physical effects. However, their influence is so subtle and so deeply entangled with the expansion rate that they remain hidden. To find them, future observations would need to break the tight link between the Hubble constant and these new terms, perhaps by combining cosmic microwave background data with independent measurements of the universe's expansion from other sources, or by probing the largest scales of the universe where the unique signature of these terms resides. For now, the standard model of a constant vacuum energy remains the best description of our universe, and these finite-temperature corrections, while theoretically present, remain beyond our reach.

Technical Summary: Kinetic-theory derivation and Bayesian analysis of the finite-temperature shifted CMB power spectrum

Problem Statement
The standard $\Lambda$ CDM model explains cosmic acceleration via a constant vacuum energy. However, finite-temperature quantum-gravitational corrections to the cosmological constant suggest the vacuum energy may not be fixed but responsive to the universe's thermal history. Previous work (Papers I and II) identified two additional density parameters, $\Omega_{\Lambda 2}$ and $\Omega_{\Lambda 3}$ , arising from these corrections. These terms scale as $a^{-4}$ and $a^{-2}$ respectively, mimicking the scaling of radiation and spatial curvature. While deterministic machine-learning analyses in prior papers located a best-fit point for these parameters, they did not address whether the data actually constrain them once all other parameter uncertainties are marginalized, nor did they elucidate the physical mechanism behind any potential degeneracies. Furthermore, the non-redundancy of these terms with standard radiation ( $\Omega_R$ ) and curvature ( $\Omega_K$ ) required further clarification, as they share the same scale-factor dependence.

Methodology
The author employs two complementary approaches to address the limitations of previous deterministic analyses:

Bayesian MCMC Analysis: The author modified the Boltzmann code class to include $\Omega_{\Lambda 2}$ and $\Omega_{\Lambda 3}$ in the background expansion rate $E(a)$ , without introducing new perturbation equations for these components. They performed a Markov Chain Monte Carlo (MCMC) analysis using the cobaya framework against the Planck 2018 likelihoods (high- $\ell$ TTTEEE, low- $\ell$ TT/EE, and lensing). This approach marginalizes over all other cosmological parameters to determine the posterior constraints on the new parameters.
Kinetic Theory Pipeline: To understand the physical imprint of these parameters, the author built a CMB power spectrum pipeline from scratch, following Weinberg's kinetic-theory formulation. Unlike standard codes that switch from a tight-coupling approximation to a full hierarchy, this pipeline integrates the full photon and polarization hierarchy directly without approximations or "switching" points. This allows for the exact evaluation of the multipole formula (Eq. 24), including the Integrated Sachs-Wolfe (ISW) term and polarization channels often dropped in analytic approximations. The pipeline is validated against class and the Planck 2018 best-fit spectrum.
Response and Fisher Analysis: Using the pipeline, the author computed the spectral response ( $\partial C_\ell / \partial \theta$ ) to $\Omega_{\Lambda 2}$ and $\Omega_{\Lambda 3}$ . They utilized a Fisher matrix analysis with the Planck covariance to determine which standard parameters can mimic (absorb) these responses.

Key Contributions and Theoretical Clarifications

Non-Redundancy: The paper clarifies that $\Omega_{\Lambda 2}$ and $\Omega_{\Lambda 3}$ are not merely relabelings of radiation and curvature. While they share the $a^{-4}$ and $a^{-2}$ scaling in the Friedmann equation, they differ fundamentally in their origin (quantum corrections to the cosmological constant vs. classical matter fields) and their perturbation behavior. Specifically, $\Omega_{\Lambda 2}$ enters only through the expansion rate $H(a)$ and does not perturb the plasma, whereas radiation does. Similarly, $\Omega_{\Lambda 3}$ enters $H(a)$ but does not modify the perturbation equations, unlike spatial curvature $\Omega_K$ , which bends the comoving-to-radial map and alters perturbation equations.
Analytic Derivation: The author derived an improved closed-form analytic expression for the CMB power spectrum (Eq. 51) by extending Weinberg's treatment. This includes retaining the polarization source term $\Pi$ in both the scattering source and the polarization channel $\bar{H}$ , leading to a polarization correction term $\epsilon_P$ that was previously dropped.
Pipeline Validation: The new pipeline was shown to reproduce the class response to the Hubble constant $h$ with high correlation (0.9965) and to match the Planck 2018 spectrum shape to within 1.34% RMS on tabulated multipoles (improving to 0.9% after correcting for a small acoustic-scale offset on a denser 31-point grid). Crucially, the pipeline demonstrated that Silk damping, Landau damping, and the ISW effect emerge naturally from the kinetic equations rather than being inserted as ad-hoc coefficients.

Results

Bayesian Constraints: The Bayesian analysis finds that both $\Omega_{\Lambda 2}$ $Ω_{Λ2}$ and $\Omega_{\Lambda 3}$ $Ω_{Λ3}$ are consistent with zero.
- $\Omega_{\Lambda 2}$ is constrained to $(-0.20 \pm 4.60) \times 10^{-7}$ . This is a two-sided null result, bounded primarily by the overall amplitude of the spectrum, which is fixed by other parameters ( $A_s$ , $\tau_{reio}$ ) and the Planck calibration prior.
- $\Omega_{\Lambda 3}$ is constrained only from one side: $\Omega_{\Lambda 3} > -0.0092$ (68% CL). The posterior does not turn over within the prior range, indicating the bound is set by the prior edge rather than the data.
Degeneracy Mechanisms: The response analysis reveals the physical reasons for these weak constraints:
- $\Omega_{\Lambda 2}$ : Its signature is nearly flat across the acoustic multipoles, acting effectively as an overall amplitude shift. Since the amplitude is already tightly constrained by standard parameters, $\Omega_{\Lambda 2}$ is degenerate with the overall normalization.
- $\Omega_{\Lambda 3}$ : Its signature is structured and sign-changing, but over the well-measured acoustic multipoles ( $\ell \gtrsim 30$ ), it is 99.9% absorbable by a shift in the Hubble constant $h$ . The small residual signal that cannot be absorbed by $h$ resides at low multipoles ( $\ell \lesssim 30$ ), where the signal is dominated by cosmic variance and the late-time ISW effect.
Reconciliation with Prior Work: The paper resolves the apparent tension between the machine-learning results of Papers I and II (which found the extended model improved the fit) and the Bayesian null result. The machine-learning analyses compared the model against a noise-free, ensemble-averaged theoretical curve ( $C_\ell(\theta_{best})$ ), where the finite-temperature parameters can smooth the fit. The Bayesian analysis compares the model against the actual noisy data with full covariance and marginalization, revealing that the "improvement" is absorbed by degeneracies with standard parameters.

Significance and Claims
The paper claims that the finite-temperature shifted cosmological constant terms are theoretically present but observationally undetected by current CMB data. The lack of detection is not due to a lack of a signature, but rather because the distinctive parts of the signature are either degenerate with the Hubble constant (at high $\ell$ ) or buried in cosmic variance (at low $\ell$ ).

The author emphasizes that their findings are modest: they do not claim to have detected new physics, nor do they propose new experiments to detect it immediately. Instead, they provide a rigorous explanation for why these parameters are currently unconstrained. They conclude that breaking the degeneracy for $\Omega_{\Lambda 3}$ would require external measurements of $h$ (e.g., from BAO or supernovae) or low- $\ell$ polarization data, while constraining $\Omega_{\Lambda 2}$ would require external amplitude information independent of the CMB temperature spectrum. The work serves as a demonstration of how kinetic-theory derivations and Bayesian marginalization can complement each other to distinguish between a "best-fit" point and a physically detectable signal.

More like this