Kick it like DESI: PNGB quintessence with a dynamically generated initial velocity

TL;DR

This paper explores a dynamical dark energy model using a PNGB axion with a non-zero initial velocity, fitting cosmological data and finding slight preference over standard models, despite theoretical tensions.

Contribution

It introduces a mechanism for sourcing initial axion velocity in quintessence models and demonstrates its compatibility with cosmological observations.

Findings

A non-zero initial velocity slightly improves fit to data.

Best-fit models involve transplanckian decay constants and flat potentials.

The scenario is in tension with string theory conjectures.

Abstract

Motivated by the hint for time-dependent dynamical dark energy from an analysis of the DESI Baryon Accoustic Oscillation (BAO) data together with information from the Cosmic Microwave Background (CMB) and Supernovae (SN), we relax the assumption of a vanishing initial velocity for a quintessence field. In particular we focus on pseudo-Nambu-Goldstone-Boson (PNGB) quintessence in the form of an axion like particle, that can arise as the phase of a complex scalar and could possess derivative couplings to fermions or topological couplings to abelian gauge fields, without upsetting the necessary flatness of its potential. We discuss mechanisms from the aforementioned interactions for sourcing an initial axion field velocity $\dot{\theta_i}$ at redshifts $3\leq z\leq 10$, that will "kick" it into motion. Driven by this initial velocity the axion will first roll up in its potential, similar to "freezing" dark energy. After it has reached the pinnacle of its trajectory, it will start to roll down, and behave as "thawing" quintessence. As a proof of concept we undertake a combined fit to BAO, SN and CMB data at the background level. We find that a scenario with $\dot{\theta_i}=\mathcal{O}(1) \; m_a$, where $m_a$ is the axion mass, is slightly preferred over both $\Lambda$CDM and the conventional "thawing" quintessence with $\dot{\theta_i}=0$. The best fit points for this case exhibit transplanckian decay constants and very flat potentials, which both are in tension with conjectures from string theory.