Phantom Mirage from Axion Dark Energy

TL;DR

This paper proposes that axions with masses near the Hubble scale can generate a phantom-like expansion history (a phantom mirage) without dark-sector couplings by behaving as dark energy at $z\sim1$ and as dark matter today. The authors show that a wide range of axion contributions can alleviate the SN–BAO distance tension similarly to thawing quintessence, but the BAO–CMB tension at high redshift persists unless additional modifications (e.g., higher optical depth $\tau$, slight curvature, or other high-$z$ physics) are invoked. They quantify improvements in fit via $\Delta\chi^2$ for representative masses (e.g., ${\log_{10}(m_{ m a})}\approx-32.9$ with $f_{ m de}\approx1$, and ${\log_{10}(m_{ m a})}\approx-32.5$ with $f_{ m de}\approx0.08$) and show that removing low-$\ell$ Planck polarization or allowing $\Omega_K>0$ can push $\Delta\chi^2$ down to values comparable to phantom dark energy models. The study introduces an emulator, AxiECAMB, to accurately predict axion-enabled observables and highlights the potential for axions to address multiple cosmological tensions with minimal new physics, while also outlining directions for future work (e.g., Early Dark Energy, reionization modeling).

Abstract

Supernova (SN) and baryon acoustic oscillation (BAO) distance measures have recently provided hints that the dark energy is not only dynamical but apparently evolves from normal to phantom dark energy between redshifts $0<z<1$. A normal axion dark energy component in the mass range just below the Hubble scale can mimic a phantom component by appearing as dark energy at $z=1$ and dark matter at $z=0$, raising the possibility of a phantom mirage. We show that there is a wide range of axion dark energy contributions that can resolve the SN-BAO tension as well as thawing quintessence does, leaving BAO tension with the cosmic microwave background (CMB) for the distance measures from $z\sim 1$ to recombination to be resolved at high redshifts. With axions, raising the optical depth to reionization to $τ\approx 0.1$ works essentially as well as $w_0-w_a$ phantom dark energy for all but the lowE CMB data, with a remaining $Δχ^2\sim -16$ compared with $Λ$CDM, whereas a small spatial curvature of $Ω_K \sim 0.003$ can largely relax the full SN-BAO-CMB tension with a total $Δχ^2 \sim -12$.

Figures (14)

• Figure 1: Phantom mirage mechanism. Top panel: true energy densities of the axion $\rho_{\rm a}$ and dark energy $\rho_{\rm de}$ compared with the effective dark energy $\rho_{\rm eff}$ that would be assumed by analyzing this case as CDM and dark energy. Bottom panel: the inferred effective equation of state $w_{\rm eff}<-1$ at high redshift despite $w_{\rm a}>-1$ approaching a CDM-like $w_{\rm a} \sim 0$ at $z=0$. Here ${\rm lg}(m_{\rm a}) =-32.5, f_{{\rm de}} = 0.1.$
• Figure 2: Likelihood profile as a function of axion mass or $\Delta\chi^2$ relative to the baseline $\Lambda$CDM model for the baseline SN+BAO+CMB analysis (lower panel). Similar improvements over $\Lambda$CDM occur in the axion phantom mirage range $-33.5 \lesssim {\rm lg}(m_{\rm a}) \lesssim -32.3$ but with very different axion contributions to the dark energy $f_{\rm de} = \Omega_{\rm a}/\Omega_{\rm de}$ (upper panel). For much smaller masses all $f_{\rm de}$ cases behave as $\Lambda$CDM with the same total $\Omega_{\rm de} = \Omega_\Lambda+\Omega_{\rm a}$ (middle panel) and for much higher masses the profile value for $f_{\rm de}$ is small and also predicts $\Lambda$CDM observables up to the accuracy of the emulator discussed in Appendix \ref{['sec:emulator_training']}.
• Figure 3: SN data compared with models in the baseline analysis. The baseline $\Lambda$CDM model fails to fit the relative magnitudes of the lowest redshift SN vs higher redshifts whereas both axion cases ${\rm lg}(m_{\rm a})=-32.9$ ($f_{\rm de}\approx 1$) and ${\rm lg}(m_{\rm a})=32.5$ ($f_{\rm de} \approx 0.08$) capture the upturn.
• Figure 4: BAO data compared with models in the baseline analysis. While the baseline $\Lambda$CDM lacks the low $z$ upturn of the axion cases, all three fit the high-$z$ BAO data comparably well. Here and in other figures "[fid]" stands for in units of the same quantity in the fiducial model. Starred datapoints represent those which are used in the likelihood whereas $D_V$ for all but the lowest redshift bin is redundant with $D_M$ and $D_H$.
• Figure 5: CMB lensing data compared with models in the baseline analysis. The baseline $\Lambda$CDM and axion cases all reduce the lensing amplitude so as to fit the BAO data reflecting tension in the BAO+CMB data.