Confronting Inflation and Reheating with Observations: Improved Predictions

Using the latest observational data, we constrain the inflationary dynamics and the subsequent reheating epoch. Predictions for both phases can be significantly improved by employing numerically computed results compared to the slow-roll approximations. These results enable a more accurate reassessment of the observational viability of inflationary models, provide tighter constraints on the reheating history, and help lift the degeneracies in the predictions of inflation and reheating dynamics. Given current observational bounds, this enables a more accurate understanding of the early universe physics.

💡 Research Summary

The paper presents a comprehensive re‑examination of inflationary dynamics and the subsequent reheating epoch using the most recent cosmological observations, notably the combined ACT‑DR6, Planck‑2018, BICEP/Keck‑2018, and DESI data (collectively referred to as P‑ACT‑LB‑BK18). The authors argue that the conventional slow‑roll approximation, which assumes potential dominance throughout inflation, breaks down near the end of inflation when kinetic energy becomes comparable to the potential. This breakdown leads to biased estimates of the number of e‑folds, the scalar spectral index (nₛ), the tensor‑to‑scalar ratio (r), and the reheating parameters (the reheating e‑fold number N_re and reheating temperature T_re).

To overcome these limitations, the authors solve the full Klein‑Gordon equation for the inflaton field together with the Friedmann equation numerically. They start from chosen initial conditions for the field value φ(N₀) and its derivative φ′(N₀) and integrate forward in the number of e‑folds N. From the numerical solution they extract the Hubble parameter H(N) and the hierarchy of Hubble‑flow parameters ε_i defined recursively as ε_{i+1}=d ln ε_i / dN (with ε₁ = -Ḣ/H²). Using next‑to‑leading order expressions for the primordial observables, namely

nₛ = 1 – 2ε₁ – ε₂ – 2ε₁² – (3+2C₁)ε₁ε₂ – C₁ε₂ε₃,

r = 16 ε₁ (1 + C₁ ε₂),

with C₁ = ln 2 + γ_E – 2 ≈ 0.7296,

they obtain predictions for nₛ and r that include higher‑order corrections absent in the simple potential‑slow‑roll formulas.

The reheating phase is modeled as an effective fluid with a constant equation‑of‑state parameter w_φ, derived from the time‑averaged pressure and energy density of an inflaton oscillating in a monomial potential V ∝ φⁿ. The analytic relation w_φ = (n‑2)/(n+2) is used as a baseline, but the authors allow a ±0.2 variation to capture possible deviations from the idealized averaging. Energy conservation across reheating yields the standard relations

ρ_re = (a_re / a_end)^{-3(1+w_φ)} ρ_end,

ρ_re = (π²/30) g_re T_re⁴,

which together give expressions for N_re and T_re in terms of the inflationary observables (nₛ, r) and model‑dependent quantities (H_k, H_end).

Two representative families of inflationary potentials are examined: (i) α‑attractors (E‑model and T‑model) with V(φ) = Λ⁴