ACT Implications for Hilltop Inflation

Technical Summary: ACT Implications for Hilltop Inflation

Problem Statement
The Atacama Cosmology Telescope (ACT) collaboration's recent DR6 data release, combined with Planck, BICEP/Keck (BK18), and DESI observations, reports a scalar spectral index of $n_{RR} = 0.974 \pm 0.003$. This represents a significant shift toward higher values compared to the Planck 2018 constraints ($n_{RR} \approx 0.965$). This shift raises critical questions regarding the viability of specific inflationary models, particularly the classes of hilltop (HI) and hilltop-squared (HSI) inflation. Previous analyses based on Planck data suggested that these models, especially at low values of the exponent $p$, could accommodate the data. However, the new spectral index constraints potentially restrict the viable parameter space, particularly concerning the inflaton parameter $\mu$ and the number of e-folds $N_*$. The paper investigates whether these models can survive the tighter constraints imposed by the ACT results without relying on restrictive approximations.

Methodology
The authors analyze a two-parameter family of inflationary potentials defined by:
$$V = \Lambda^4 \left(1 - \left(\frac{\phi}{\mu}\right)^p\right)^n$$
where $\Lambda$ is the energy scale, $\mu$ is the inflaton parameter, $p$ is an exponent, and $n$ distinguishes the model classes: $n=1$ for hilltop inflation (HI) and $n=2$ for hilltop-squared inflation (HSI).

The study employs the following methodological steps:

1. Analytical Derivation: The authors derive the spectral index $n_{RR}$ and the tensor-to-scalar ratio $r$ using slow-roll parameters $\epsilon_V$ and $\eta_V$. They examine these observables under three regimes:
   • General slow-roll approximation.
   • Small-field approximation (SF.II), where $\phi \ll \mu$.
   • Small-field with sub-Planckian $\mu$ (SF.III), where $\mu \ll M_{Pl}$.
2. Constraint Application: The models are tested against the combined P-ACT-LB-BK18 DR6 dataset. The analysis compares these new constraints against previous Planck-only constraints ($n_{RR} = 0.965 \pm 0.005$, $r \leq 0.035$) and the standard e-fold range $N_* \in [50, 70]$.
3. Parameter Scanning: The authors perform systematic scans over the parameter $\mu$ (in steps of one Planck mass) and the exponent $p$ (specifically $p=4, 7, 10, 20$). They explicitly avoid dropping the $\mu$-dependence often assumed in simplified scaling relations (like $n_{RR} = 1 - 2(p-1)/((p-2)N)$) to determine the full viability of the models.
4. Reheating Analysis: The study incorporates constraints from the reheating epoch, calculating the number of reheating e-folds $N_{rh}$ to determine if matter-dominated reheating is viable within the allowed parameter spaces.

Key Contributions and Results

• Rejection of Sub-Planckian Regimes: The most significant finding is that for both HI and HSI models, the new ACT constraints exclude the regime of sub-Planckian $\mu$ ($\mu \ll M_{Pl}$). In the SF.III regime (small field and small $\mu$), the spectral index is predicted to be $n_{RR} \approx 1 - 3/N_*$, which falls outside the $2\sigma$ confidence region of the ACT data. Consequently, viable models must now operate with super-Planckian values of $\mu$.
• Differential Impact on HSI vs. HI:
  • Hilltop-Squared (HSI, $n=2$): These models face severe constraints. For low $p$ (e.g., $p=4$), the models cannot cover the full range of the spectral index allowed by ACT data. Furthermore, no viable HSI model allows for matter-dominated reheating. The viable parameter space for HSI shifts entirely to higher $\mu$ values, and the number of e-folds required to match the data generally exceeds the canonical $N_* \in [50, 60]$ range.
  • Hilltop (HI, $n=1$): While also constrained, the impact is milder. HI models can realize the full spectral index range allowed by ACT data. Crucially, for every $p$, there exists a region in parameter space where matter-dominated reheating is viable, corresponding to $N_* \in [50, 60]$.
• Scaling Relations and Stratification: The paper highlights that the spectral index and tensor ratio exhibit a "stratification" dependent on $\mu$. The simple, universal scaling relations derived in the SF.III regime (independent of $\mu$ and $n$) are shown to be insufficient. The full parameter dependence reveals that the spectral tilt $\delta = 1 - n_{RR}$ has a significant energy dependence on both $N$ and $\mu$.
• Exponent Dependence: Low-$p$ models (e.g., $p=3$) are almost excluded for HSI and severely constrained for HI. Models with higher $p$ ($p \geq 4$) remain viable but require super-Planckian $\mu$. As $p$ increases, the SF.III regime ranges eventually overlap with the $2\sigma$ likelihood region, but the viable domain remains shifted to higher $\mu$.

Significance and Claims
The paper claims that the shift in the spectral index reported by ACT fundamentally alters the viability landscape for hilltop and hilltop-squared inflation. The primary significance lies in demonstrating that:

1. Approximations are insufficient: Relying on the small-field, sub-Planckian approximation (SF.III) leads to the incorrect conclusion that these models are entirely excluded. A full analysis retaining $\mu$-dependence is necessary to identify viable super-Planckian regimes.
2. Model Discrimination: The new data creates a sharper distinction between HI and HSI. While HI models retain regions compatible with standard reheating and canonical e-fold numbers, HSI models are pushed into a regime where matter-dominated reheating is impossible and e-fold numbers are higher.
3. Parameter Space Reduction: The viable parameter space for these theories is significantly reduced compared to Planck-era results. Specifically, the sub-Planckian regime is no longer accessible for these models under ACT constraints.

The authors conclude that while low-$p$ HSI models are nearly excluded, models with $p \geq 4$ can still satisfy the CMB constraints provided $\mu$ is super-Planckian. They note that future constraints from the Simons Observatory could further restrict the lower $p$ regime if gravitational waves are not detected.