Cosmology of axion dark energy in supersymmetric models and constraints on high scale parameters

TL;DR

This work analyzes a dynamical dark energy scenario where an ultralight axion-like field with a multi-cosine pNGB potential drives cosmic acceleration and interacts feebly with dark matter. A Lagrangian QCDM framework is developed and implemented in a Boltzmann solver with a reformulation that tames rapid dark matter oscillations, enabling robust background and perturbation evolution. Using Planck, DESI-DR2 BAO, and SN datasets, the authors constrain the axion decay constant to sub-Planckian values and require a feeble DM-DE coupling, with N=2 often yielding a better fit than N=1. Higher-N cases (N=3,4) exhibit transmutation of thawing to freezing quintessence due to the potential alone, while LCDM remains preferred by information criteria in many dataset combinations; the study thereby links high-scale axion physics from SUSY/String models to cosmological observables and informs the potential resolution of the Hubble tension.

Abstract

An analysis is given of interacting dark energy and dark matter where the dark energy is assumed to be an ultralight axionic field with a pseudo-Nambu-Goldstone Boson potential which is in general a superposition of $N$ number of cosine terms motivated by supergravity and string models with a $U(1)$ global symmetry, where the symmetry is broken by instanton effects. The case $N=2$ is investigated in detail and a fit to cosmological data is performed where it is found that a better fit is obtained in comparison with the $N=1$ case. The fits also constrain high scale parameters, i.e., the axion decay constant which is determined to be sub-Planckian, a result consistent with string theory that disfavors trans-Planckian axion decay constant. Furthermore, the dark energy-dark matter interaction strength is constrained to be feeble, i.e., $λ\lesssim 4\times 10^{-6}$ m$_{\rm Pl}^{-2}$ Mpc$^{-2}$. We study possible implications of this type of potential on the Hubble tension and on the dynamics of the dark energy equation of state using the DESI-DR2 data. For the cases $N=3,4$, the analysis exhibits the phenomenon of transmutation even in the absence of coupling to dark matter, where thawing quintessence transmutes to freezing quintessence. The analysis is internally consistent in its treatment of the dark energy-dark matter interaction as it is based on an underlying Lagrangian, in contrast with several previous works where the sources are chosen in an ad hoc manner to satisfy energy conservation.

Figures (8)

• Figure 1: The matter power spectrum versus the wavenumber $k$ (left panel) and the temperature TT power spectrum as a function of the multipoles (right panel) for three cases: $N=1$, $N=2$ with equal coefficients ($c_1=c_2=1$ with $\log\mu^4=-7.0$) and $N=2$ with different coefficients ($c_1=2.0$, $c_2=-0.1$ and $\log\mu^4=-7.0$). No DM-DE interaction is present. The $\Lambda$CDM case is shown as a dashed black curve.
• Figure 2: Top panels: The dark energy potential $V_2(\phi)$ plotted against the redshift (left panel) and against the field $\phi$ (right panel) for the cases $N=1,2,3,4$ with equal coefficients, $c_i=1$ ($i=1,\cdots,4)$, and with $\log\mu^4=-7.0$. Bottom panels: the matter (dotted) and DE (solid) densities (left panel) and the DE equation of state (right panel) plotted against the redshift. No DM-DE interaction is present.
• Figure 3: The 1D and 2D marginalized posteriors for the $N=1$ (top left panel) and $N=2$ (top right panel) cases, showing correlations between different cosmological parameters for the three data sets. Here $c_1=c_2=1$. Bottom panel: same as top right panel but for $c_1\neq c_2$ and $\log\mu^4=-7.0$.
• Figure 4: Results for the posterior distributions of $w_0$ and $w_a$ for three data sets pertaining to the $N=1$ case. The best fit for each data set and DESI's contours are also shown.
• Figure 5: Results for the posterior distributions of $w_0$ and $w_a$ for three data sets pertaining to the cases $N=2$ with equal coefficients ($c_1=c_2=1$) and $N=2$ with different coefficients ($c_1\neq c_2$). The best fit for each data set and DESI's contours are also shown.