Neutrinos help reconcile Planck measurements with both Early and Local Universe

TL;DR

The paper addresses a tension between the large tensor-to-scalar ratio suggested by BICEP2 and the Planck ΛCDM model, along with independent local-Universe tensions in $H_0$ and cluster abundances. It tests extensions with a sterile neutrino by introducing $ abla N_{ m eff}$ and $m_s$ within $ u ext{ΛCDM}$ and $ u r ext{ΛCDM}$, analyzing data categorized into early (EC) and local (CL) tensions and their combination (ECL) using CosmoMC. The main finding is strong joint evidence for a sterile neutrino, with $ abla N_{ m eff}=0.98\pm0.26$ and $m_s=0.52\pm0.13$ eV under the full ECL data, corresponding to ~3.8σ and ~4σ significance, and a compatible $H_0$ around $0.74\pm0.04$ along with an adjusted growth parameter $S_8$. These results also imply inflationary scenarios with $n_s\approx1$ and $r\approx0.2$, suggesting a potentially new cosmological standard model if systematics in clusters and high-$\ell$ CMB data hold; future observations will decisively test this neutrino-based resolution.

Abstract

In light of the recent BICEP2 B-mode polarization detection, which implies a large inflationary tensor-to-scalar ratio r_{0.05}=0.2^{+0.07}_{-0.05}, we re-examine the evidence for an extra sterile massive neutrino, originally invoked to account for the tension between the cosmic microwave background (CMB) temperature power spectrum and local measurements of the expansion rate H0 and cosmological structure. With only the standard active neutrinos and power-law scalar spectra, this detection is in tension with the upper limit of r<0.11 (95% confidence) from the lack of a corresponding low multipole excess in the temperature anisotropy from gravitational waves. An extra sterile species with the same energy density as is needed to reconcile the CMB data with H0 measurements can also alleviate this new tension. By combining data from the Planck and ACT/SPT temperature spectra, WMAP9 polarization, H_0, baryon acoustic oscillation and local cluster abundance measurements with BICEP2 data, we find the joint evidence for a sterile massive neutrino increases to DeltaNeff=0.98\pm 0.26 for the effective number and ms= 0.52\pm 0.13 eV for the effective mass or 3.8 sigma and 4 sigma evidence respectively. We caution the reader that these results correspond to a joint statistical evidence and, in addition, astrophysical systematic errors in the clusters and H0 measurements, and small-scale CMB data could weaken our conclusions.

Figures (9)

• Figure 1: BICEP2 measurement of the tensor-scalar ratio $r_{0.05}$ (bands) compared with the posterior probability distribution of the C data set in the $r\Lambda$CDM model space (black curve) and $\nu r\Lambda$CDM space (red curve). In the former, the measurement is in strong tension with the posteriors whereas the addition of massive sterile neutrinos in the latter allows high $r$. For both the curves and the band, the tensor-to-scalar ratio is evaluated at a pivot scale of $k=0.05$Mpc$^{-1}$ unlike elsewhere.
• Figure 2: Early Universe tension and neutrinos. In the $\nu r\Lambda$CDM parameter space the EC data set favors $\Delta N_{\rm eff}$$>0$ in order to offset the excess large angle temperature anisotropy implied by the high tensor-scalar ratio $r$ (68%, 95% contours here and below). This in turn is driven by the degeneracy between $\Delta N_{\rm eff}$ and $n_s$ illustrated in Fig. \ref{['fig:tilt']}. In brief, gravitational waves add power at low $\ell$, requiring larger $n_s$ to compensate. Larger $n_s$ then requires larger $\Delta N_{\rm eff}$ to agree with the higher-$\ell$ CMB.
• Figure 3: In the $\nu r\Lambda$CDM parameter space the EC data set allows a positive change in the tilt when $\Delta N_{\rm eff}$ is increased explaining the mechanism by which the large angle temperature anisotropy is reduced.
• Figure 4: By allowing for tensors and neutrinos, the EC data set in the $\nu r\Lambda$CDM model favors higher values for $H_0$. Note that the actual measurements of $H_0$ are not imposed here as a prior -- the BICEP2 central value of $r$ in $\nu r\Lambda$CDM predicts an $H_0$ in concordance with observations.
• Figure 5: $H_0$ and $S_8$ posterior probability distributions in the $\nu r\Lambda$CDM parameter space using early Universe (EC) data alone compared with the late Universe measurements. To emphasize, the black 1D posteriors plotted here have been derived without any use of local Universe data. The addition of neutrinos and tensors makes the $H_0$ posterior fully compatible with measurements (bands) and $S_8$ substantially more compatible, though some residual tension remains. The dotted line represents the shift in the central value of the cluster abundance measurements under the assumption of a 9% systematic increase in cluster masses.