Primordial full bispectra from the general bounce cosmology

TL;DR

This paper calculates the full bispectra of scalar and tensor perturbations in a general bounce cosmology, revealing differences in non-Gaussianity shapes between two model branches and identifying parameter spaces consistent with observations.

Contribution

It introduces a comprehensive analysis of primordial bispectra in a general bounce model, comparing two distinct branches and their non-Gaussian signatures, including consistency with Maldacena's relation.

Findings

Differences in bispectrum shapes between the two bounce branches.

Reproduction of Maldacena's consistency relation in one branch.

Identification of parameter space satisfying observational constraints.

Abstract

Primordial non-Gaussianities are key quantities to test early universe scenarios. In this paper, we compute full bispectra of scalar and tensor perturbations generated during a contracting phase in a general bounce model. The general bounce model consists of two branches: one realizes scale-invariant scalar and tensor power spectra from perturbations whose amplitudes become constant on superhorizon scales, as in de Sitter inflation, while the other realizes scale-invariant power spectra from perturbations whose amplitudes grow on superhorizon scales, as in matter bounce cosmology. We study the auto- and cross-bispectra originating from the scale-invariant scalar and tensor perturbations in these two branches. We investigate the amplitudes and shapes of non-Gaussianities and find that the differences between the two branches manifest for equilateral and squeezed momentum triangle configurations. In particular, one of the branches in which the superhorizon perturbations are conserved reproduces the so-called Maldacena's consistency relation. By examining perturbativity conditions and considering current observational constraints on primordial non-Gaussianities, we also find a viable parameter space in which both theoretical and observational constraints are satisfied simultaneously.