Cosmological observables, IR growth of fluctuations, and scale-dependent anisotropies

TL;DR

The paper analyzes how infrared growth of fluctuations during inflation affects cosmological observables and proposes an IR-safe framework that separates local observables from global IR effects. By introducing a scale-dependent metric and a scale-dependent physical momentum, it shows that leading IR corrections can be resummed into an effective spectrum P_0(κ), reducing sensitivity to an artificial IR cutoff. For late-time observers, it argues that long-wavelength modes distort but can be scaled out in physical coordinates, with the observer’s horizon providing the IR cutoff and potentially yielding small but detectable statistical anisotropies, especially in 21 cm measurements. The work connects these IR-resummation ideas to renormalization-group concepts and discusses implications for the breakdown of perturbation theory at very late times and for broader inflationary scenarios.

Abstract

We extend semiclassical methods in inflationary cosmology that capture leading IR corrections to correlators. Such large IR effects can be absorbed into a coordinate change when examining sufficiently local observables, but not when comparing observations at large separation in scales, such as seen by a late-time observer. The analysis is facilitated by definition of a scale-dependent metric and physical momentum. These assist definition of "IR-safe" observables seen by a post-inflationary observer, which are contrasted to those based on the local geometry of the reheating surface. For such observables, the observer's horizon provides an effective IR cutoff. IR growth contributes to enhanced statistical inhomogeneities/anisotropies at short scales, observation of which by a present day observer might be sought in 21 cm measurements. Such IR corrections are argued to grow large for a very late-time observer.