Current constraints on cosmological scenarios with very low reheating temperatures

TL;DR

This work analyzes cosmological scenarios with very low reheating temperatures around the MeV scale, where neutrino thermalization is incomplete and both BBN and CMB observables are altered. It advances the field by computing momentum-dependent neutrino distributions with FortEPiaNO, updating primordial yields with a consistent PArthENoPE treatment, and performing a joint Planck+lensing+DESI (plus BAO) plus BBN analysis while carefully comparing prior choices. The principal finding is a robust 95% CL lower bound of $T_ ext{RH} > 5.96$ MeV when including light-element abundances, Planck data, and BAO, making this the most stringent constraint to date on very low reheating scenarios. The study also clarifies how $T_ ext{RH}$ impacts $oldsymbol{ heta}_{ ext{LCDM}}$ and $ otal{m_ u}$, and provides forecasts showing next-generation CMB experiments will further tighten these limits via $N_ ext{eff}$ and $Y_p$ measurements, with implications for models featuring late-time reheating or additional relativistic species.

Abstract

We present a comprehensive analysis of the effects of models with very low reheating scenarios ($T_\text{RH} \sim \mathcal{O}(\text{MeV})$) on the cosmological observables and derive corresponding bounds on the reheating temperature. With respect to previous work, our study includes a more precise computation of neutrino distribution functions, leveraging the latest datasets from cosmological surveys. We perform a joint analysis that combines constraints from Big Bang Nucleosynthesis, the Cosmic Microwave Background, and galaxy surveys, alongside separate investigations of these datasets, carefully assessing the impact of different choices of priors. At the $95\%$ confidence level, we establish a lower bound on the reheating temperature of $T_\text{RH} > 5.96 \; \text{MeV} $, representing the most stringent constraint to date.

Figures (12)

• Figure 1: Final neutrino energy density expressed in terms of $N_\mathrm{eff}$, as a function of the reheating temperature. The horizontal line indicates the standard value, $N_\mathrm{eff}=3.044$. Solid lines indicate our new results with (blue) and without (orange) flavor neutrino oscillations. For comparison, the results of the analysis of deSalas:2015glj are shown with a dotted red line. Filled regions indicate the present bounds from Planck Planck:2019nip and the future sensitivity from the Simons observatory Ade:2018sbj on $N_\mathrm{eff}$ at 95% CL.
• Figure 2: Contour plots for the pure BBN likelihood function (arbitrary units) in the plane $T_\mathrm{RH}-\omega_\mathrm{b}$.
• Figure 3: Contour plots for the Planck+lensing+DESI marginalized likelihood function (arbitrary units). The first panel shows the $T_\mathrm{RH}-\omega_b$ plane as a comparison with \ref{['fig:BBN_like']}. The second and third panels show two key degenerations between parameters of the model.
• Figure S1: Deuterium (left) and helium (right) abundances as a function of the reheating temperature for $\Omega_b h^2=0.02242$ (see the text for a detailed discussion). Shaded cyan regions correspond to the $68\%$ and $95\%$ CL regions recommended in ParticleDataGroup:2022pth, while shaded green regions correspond to $68\%$ and $95\%$ theoretical errors.
• Figure S2: Comparison between the approximated (FD and FD+$Y_p$, dashed and dotted lines respectively, in all plots) and exact (PSD+$Y_p$, solid) computation of CMB temperature power spectra. The approximated implementations use a Fermi-Dirac distribution for the neutrinos inside the Boltzmann code CLASS, while the exact one accounts for the actual form derived from FortEPiaNO, thus including spectral distortions. In the FD implementation the helium abundance is computed accounting only for changes in the expansion history in the radiation-dominated era. In the FD+$Y_p$ and PSD+$Y_p$ implementations this is instead computed also taking into account the nonthermal neutrino distribution. See text for details. Left panel: Relative difference between spectra computed using the FD or FD+$Y_p$ and the PSD+$Y_p$ implementation, for different values of the reheating temperature. Middle panel: Difference between spectra for different values of the reheating temperature and the Planck 2018 best-fit model, using the three implementations under considerations. Also shown are the Planck 2018 data with the corresponding uncertainty. The grey shaded area shows the uncertainty of a cosmic-variance limited experiment observing 70% of the sky, with the same multipole binning $\Delta \ell = 30$ as the Planck data. Right panel: Difference between spectra computed using the FD or FD+$Y_p$ and the PSD+$Y_p$ implementation, for different values of the reheating temperature. In this plot, the theoretical spectra have been binned following the same scheme as the Planck data. The shaded regions show the observational uncertainty of Planck and of the same cosmic-variance limited experiment as in the middle panel.