Reheating, Multifield Inflation and the Fate of the Primordial Observables

TL;DR

This paper investigates how the reheating process after two-field inflation influences primordial observables like non-Gaussianity and spectral index, highlighting the importance of reheating physics in interpreting cosmological data.

Contribution

It provides a numerical analysis of reheating effects on curvature perturbations and demonstrates the sensitivity of non-Gaussianity to reheating details in two-field inflation models.

Findings

f_NL can remain non-zero if initial non-Gaussianity exists.

f_NL is sensitive to reheating timescale, especially in certain potentials.

n_s remains relatively insensitive to decay rate variations.

Abstract

We study the effects of perturbative reheating on the evolution of the curvature perturbation \zeta, in two-field inflation models. We use numerical methods to explore the sensitivity of f_NL, n_s and r to the reheating process, and present simple qualitative arguments to explain our results. In general, if a large non-Gaussian signal exists at the start of reheating, it will remain non zero at the end of reheating. Unless all isocurvature modes have completely decayed before the start of reheating, we find that the non-linearity parameter, f_NL, can be sensitive to the reheating timescale, and that this dependence is most appreciable for `runaway' inflationary potentials that only have a minimum in one direction. For potentials with a minimum in both directions, f_NL can also be sensitive to reheating if a mild hierarchy exists between the decay rates of each field. Within the class of models studied, we find that the spectral index n_s, is fairly insensitive to large changes in the field decay rates, indicating that n_s is a more robust inflationary observable, unlike the non-linearity parameter f_NL. Our results imply that the statistics of \zeta, especially f_NL, can only be reliably used to discriminate between models of two-field inflation if the physics of reheating are properly accounted for.