Eternal inflation near inflection points: a challenge to primordial black hole models

Technical Summary: Eternal inflation near inflection points: a challenge to primordial black hole models

Problem Statement
Primordial black hole (PBH) models based on inflection point potentials are a popular mechanism for generating the large density perturbations required for PBH formation. These models typically feature a potential with a local minimum followed by a local maximum near the end of inflation. While the classical trajectory allows the inflaton to roll over this feature, stochastic quantum noise can oppose the classical drift. In rare patches, this noise can delay reheating indefinitely, leading to eternal inflation. The paper addresses the question of whether such inflection point models inevitably undergo eternal inflation and, if so, what the consequences are for the compatibility of these models with cosmological observations, specifically the Cosmic Microwave Background (CMB) and large-scale structure.

Methodology
The authors employ the stochastic inflation formalism, where the inflaton field $\phi$ evolves according to a Fokker–Planck equation. The probability distribution $P(\phi, N)$ of the field over the number of e-folds $N$ is governed by a drift term (classical rolling) and a diffusion term (quantum noise).

1. Eternal Inflation Criterion: The paper establishes that eternal inflation occurs if the expected physical volume of inflating space does not vanish as $N \to \infty$. This condition is rigorously linked to the lowest eigenvalue $\lambda_1$ of the Fokker–Planck operator $L_{FP}$. Specifically, eternal inflation is guaranteed if $\lambda_1 \le 3$.
2. Sub-potential Analysis: Instead of solving the Fokker–Planck equation for the full, complex potential (which is numerically unstable due to disparate scales), the authors analyze restricted "sub-potentials" around the local maximum and minimum of the inflection point feature. They prove that if a sub-potential satisfies $\tilde{\lambda}_1 \le 3$, the full potential also undergoes eternal inflation.
3. Analytical Solutions: The authors derive analytical solutions for $\lambda_1$ in two simplified regimes:
   • Constant Drift: Corresponding to a linear potential.
   • Linear Drift: Corresponding to a quadratic potential (parabolic approximation near extrema). This involves solving Kummer's equation using confluent hypergeometric functions.
4. Model Testing: These analytical results are applied to three specific PBH models from the literature (Models I, II, and III), which feature polynomial, $\alpha$-attractor, and KKLT-based potentials, respectively.

Key Contributions and Results

• Sufficient Condition for Eternal Inflation: The paper derives a sufficient condition for eternal inflation in quadratic regions of the potential. For a wide quadratic region (where the field range $\phi_b \gg H$), eternal inflation occurs if the second slow-roll parameter satisfies $\eta_V \ge -6$ (or equivalently $\eta_H \ge -3$).
• Application to PBH Models: When applied to the three example models, the authors find that all satisfy the condition for eternal inflation.
  • Around the local maximum (hilltop), the models typically have $\eta_V \approx -0.4$ to $-0.5$, well within the range $\eta_V \ge -6$.
  • Around the local minimum, the models have $\eta_V > 0$, which guarantees eternal inflation in the wide-field limit ($\phi_b \gg H$) because the probability distribution settles into a static equilibrium ($\lambda_1 \to 0$).
  • The calculated eigenvalues $\lambda_1$ for all models are significantly less than 3 (ranging from $\sim 0.3$ to $0.5$), confirming that these models inflate eternally.
• Global Structure and "Baby Universes": The authors argue that eternally inflating regions in these models correspond to "Type II" perturbations. These regions form inside primordial black holes, hidden behind event horizons. Inside these black holes, inflation continues indefinitely, creating "baby universes" that are causally disconnected from the reheated exterior.
• Measure Problem and Observational Incompatibility: The paper argues that if volume weighting is used to assign probabilities (a standard approach in eternal inflation), the total volume of the universe is dominated by these inhomogeneous baby universes (specifically Type U2 and U3 regions).
  • U2 Universes: Formed by fields exiting the inflection point downwards; they are highly inhomogeneous and likely collapse quickly.
  • U3 Universes: Formed by fields kicked back up the potential; while they may resemble our universe locally, they possess a "maximal distance of homogeneity" that is often smaller than the observable universe.
  • Conclusion: Under volume weighting, the "typical" observer would find themselves in a highly inhomogeneous universe incompatible with CMB observations. Thus, eternal inflation renders typical inflection point PBH models incompatible with large-scale structure observations.

Significance and Claims
The paper claims that eternal inflation is a generic and difficult-to-avoid feature of single-field inflection point PBH models. The authors do not claim to have ruled out these models definitively but argue that their viability is severely challenged if one accepts the standard volume-weighting measure.

• Challenge to Viability: The existence of a dominant, inhomogeneous volume of baby universes suggests that these models cannot naturally explain the observed homogeneity of the universe without fine-tuning or rejecting the volume-weighting measure.
• Methodological Tool: The paper provides a practical tool for model builders: by calculating $\lambda_1$ for the quadratic approximation of the potential near extrema, one can quickly determine if a proposed PBH model suffers from eternal inflation.
• Limitations: The authors acknowledge that the conclusion depends on the choice of measure. If volume weighting is rejected, the models might still be viable, but this would undermine the motivation for inflation as a generic initial condition. They also note that avoiding eternal inflation would require either very sharp features (small $\phi_b/\sigma$) or very steep potentials (large $|\eta_V|$), which may conflict with the requirement to produce a significant abundance of PBHs in the desired mass window.

In summary, the paper demonstrates that the very features (flatness near inflection points) that make these models attractive for PBH production also make them prone to eternal inflation, potentially rendering them inconsistent with cosmological observations under standard probabilistic interpretations.