Can Early Dark Energy Explain EDGES?

TL;DR

The paper tackles the EDGES 21 cm absorption feature at $z\sim15$--$20$, which is deeper than standard $\Lambda$CDM expectations, by testing whether an early dark energy (EDE) component with $w=-1$ at early times and rapid decay at a critical redshift $z_c$ can drive earlier gas–photon decoupling. The authors formulate the EDE energy density as $\rho_{ee}(a)=\rho_c \Omega_{ee} \dfrac{1+a_c^6}{a^6+a_c^6}$ with $a_c=1/(1+z_c)$ and $p_{ee}(a)=\rho_{ee}(a)\dfrac{a^6-a_c^6}{a^6+a_c^6}$, and compute the thermal history using a modified Recfast to evaluate the gas temperature $T_{gas}$ and radiation temperature $T_{\gamma}$. They find that, for representative $(\Omega_{ee}, z_c)$, the decoupling occurs earlier due to an enhanced expansion rate $H(z)$, yielding $T_{gas}(z=20) \approx 6.5$–$8.3$ K, which partially approaches the EDGES requirement. However, the parameter space that would fully explain the EDGES depth is strongly constrained by Planck CMB TT power spectrum measurements, the angular-scale distance $D_{SLS}$, and the local $H_0$ tension, and would require lowering $H_0$ in tension with observed values (and potentially adding extra relativistic species that worsen damping-tail constraints). The combined constraints indicate that non-finely-tuned background modifications such as EDE are unlikely to reconcile EDGES with the full set of cosmological observations.

Abstract

The Experiment to Detect the Global Epoch of Reionization Signature (EDGES) collaboration has reported the detection of an absorption feature in the sky-averaged spectrum at $\approx 78$ MHz. This signal has been interpreted as the absorption of cosmic microwave background (CMB) photons at redshifts $15 \lesssim z \lesssim 20$ by the 21cm hyperfine transition of neutral hydrogen, whose temperature is expected to be coupled to the gas temperature by the Wouthuysen-Field effect during this epoch. Because the gas is colder than the CMB, the 21cm signal is seen in absorption. However, the absorption depth reported by EDGES is more than twice the maximal value expected in the standard cosmological model, at $\approx 3.8σ$ significance. Here, we propose an explanation for this depth based on "early dark energy" (EDE), a scenario in which an additional component with equation of state $w=-1$ contributes to the cosmological energy density at early times, before decaying rapidly at a critical redshift, $z_c$. For $20 \lesssim z_c \lesssim 1000$, the accelerated expansion due to the EDE can produce an earlier decoupling of the gas temperature from the radiation temperature than that in the standard model, giving the gas additional time to cool adiabatically before the first luminous sources form. We show that the EDE scenario can successfully explain the large amplitude of the EDGES signal. However, such models are strongly ruled out by observations of the CMB temperature power spectrum. Moreover, the EDE models needed to explain the EDGES signal exacerbate the current tension in low- and high-redshift measurements of the Hubble constant. We conclude that non-finely-tuned modifications of the background cosmology are unlikely to explain the EDGES signal while remaining consistent with other cosmological observations.

Figures (3)

• Figure 1: Expansion rate (solid curves) and Compton-heating rate (dashed curves) for standard $\Lambda$CDM (blue solid/green dashed) and various EDE models. The red solid/magenta dashed curves show the rates for $\Omega_{ee}=2\times 10^{-5}$ and $z_c=100$; the orange solid/cyan dashed curves show the rates for $\Omega_{ee}=2\times 10^{-5}$ and $z_c=50$; and the yellow solid/pink dashed curves show the rates for $\Omega_{ee}=4\times 10^{-5}$ and $z_c=100$. The decoupling of the gas temperature from the radiation temperature occurs when $H(z) \approx 1/t_C(z)$, i.e., when Compton-heating falls out of equilibrium. In the EDE models, decoupling occurs earlier, and thus the gas temperature is lower at late times than it is in $\Lambda$CDM.
• Figure 2: Gas temperature (dashed curves) and radiation temperature (solid blue curve) for standard $\Lambda$CDM and various EDE models. The radiation temperature evolution is identical in all models, and thus we show it only in $\Lambda$CDM for clarity. The dashed green curve shows the gas temperature evolution for $\Lambda$CDM; the dashed magenta curve shows the same quantity for an EDE model with $\Omega_{ee}=2\times 10^{-5}$ and $z_c=100$; the dashed cyan curve shows the same quantity for an EDE model with $\Omega_{ee}=2\times 10^{-5}$ and $z_c=50$; and the dashed pink curve shows the same quantity for an EDE model with $\Omega_{ee}=4\times 10^{-5}$ and $z_c=100$. The black point indicates the gas temperature at $z\approx 20$ implied by the EDGES signal, assuming the standard CMB radiation background. In the EDE models shown here, the gas temperature is lower at this redshift than in $\Lambda$CDM, due to the earlier decoupling of the gas from the radiation.
• Figure 3: Constraints in the EDE $(\Omega_{ee}, z_{c})$ parameter space from cosmological observables compared to the region necessary to explain the EDGES signal. The region above the red curve is excluded by Planck measurements of the CMB TT power spectrum KK2016 ($1\sigma$ exclusion), primarily because of the large ISW contribution associated with the EDE epoch. The region above the black curve is excluded by purely geometric constraints from the CMB (i.e., the angular size of the acoustic scale) when combined with local measurements of $H_0$ and the DESy1 constraint on $\Omega_m$. Given concerns about tension in measurements of $H_0$, we adopt a conservative $4\sigma$ exclusion limit for the black curve. The region above the blue curve contains the models necessary to explain the EDGES signal at $2\sigma$ by lowering the gas temperature at $z \approx 20$ compared to that in $\Lambda$CDM. The color bar indicates the gas temperature at $z=20$ for each point in parameter space. The necessary models are strongly excluded (note the logarithmic scales on the axes).