Scale-dependent non-Gaussianity and the CMB Power Asymmetry

TL;DR

This paper proposes a new way to parametrize the scale dependence of non-Gaussianity in the CMB, addressing limitations of previous models and exploring implications for observed power asymmetry.

Contribution

It introduces an alternative parametrisation for $f_{NL}$ that remains valid when it changes sign, motivated by the self-interacting curvaton scenario, and applies it to CMB power asymmetry analysis.

Findings

Strong scale dependence of $f_{NL}$ can lead to large $g_{NL}$ and quadrupolar asymmetry.

Model regimes with significant scale dependence are constrained by observations.

The new parametrisation improves modeling of non-Gaussianity across scales.

Abstract

We introduce an alternative parametrisation for the scale dependence of the non-linearity parameter $f_{\rm NL}$ in quasi-local models of non-Gaussianity. Our parametrisation remains valid when $f_{\rm NL}$ changes sign, unlike the commonly adopted power law ansatz $f_{\rm NL}(k) \propto k^{ n_{f_{\rm NL}} }$. We motivate our alternative parametrisation by appealing to the self-interacting curvaton scenario, and as an application, we apply it to the CMB power asymmetry. Explaining the power asymmetry requires a strongly scale dependent non-Gaussianity. We show that regimes of model parameter space where $f_{\rm NL}$ is strongly scale dependent are typically associated with a large $g_{\rm NL}$ and quadrupolar power asymmetry, which can be ruled out by existing observational constraints.