Gauge field production in SUGRA inflation: local non-Gaussianity and primordial black holes

TL;DR

This paper explores how gauge field production during supergravity inflation can lead to observable non-Gaussianities and primordial black holes, with mechanisms depending on gauge field mass and coupling strength.

Contribution

It demonstrates that gauge field production in supergravity inflation naturally produces significant non-Gaussianity and black hole constraints, expanding the understanding of inflationary observational signatures.

Findings

Large non-Gaussianity (f_NL=O(10)) can be generated in supergravity models.

Gauge fields can acquire mass through scalar couplings, affecting perturbation outcomes.

Primordial black hole bounds constrain gauge coupling parameters.

Abstract

When inflation is driven by a pseudo-scalar field \chi coupled to vectors as \alpha/4 \chi F \tilde F, this coupling may lead to a copious production of gauge quanta, which in turns induces non-Gaussian and non-scale invariant corrections to curvature perturbations. We point out that this mechanism is generically at work in a broad class of inflationary models in supergravity hence providing them with a rich set of observational predictions. When the gauge fields are massless, significant effects on CMB scales emerge only for relatively large \alpha. We show that in this regime, the curvature perturbations produced at the last stages of inflation have a relatively large amplitude that is of the order of the upper bound set by the possible production of primordial black holes by non-Gaussian perturbations. On the other hand, within the supergravity framework described in our paper, the gauge fields can often acquire a mass through a coupling to additional light scalar fields. Perturbations of these fields modulate the duration of inflation, which serves as a source for non-Gaussian perturbations of the metric. In this regime, the bounds from primordial black holes are parametrically satisfied and non-Gaussianity of the local type can be generated at the observationally interesting level f_NL =O(10).