Combining Planck with Large Scale Structure gives strong neutrino mass constraint

TL;DR

This work addresses the cosmological constraint on the sum of neutrino masses, $\sum m_\nu$, by combining Planck CMB data with large-scale structure information from the WiggleZ survey and BAO measurements. It systematically explores several neutrino scenarios (degenerate, normal/inverted hierarchies, and sterile species) within a $\Lambda$CDM framework, including cases with varying $N_\mathrm{eff}$, and uses MCMC sampling to derive 95% CL upper limits while accounting for priors. The key result is the strongest cosmological limit to date, $\sum m_\nu<0.18$ eV (95% CL) for certain priors, with the constraint tightening most for the $1+2\nu$ scenario; allowing $N_\mathrm{eff}$ to vary softens the bound to $\lesssim0.37$ eV, and hints of extra relativistic species emerge at around the 1σ level when combined with WiggleZ. The findings highlight the potential of upcoming large-scale structure surveys to probe neutrino mass hierarchies, while emphasising the dependence on modelling of non-linear scales and the need for robust theoretical frameworks to fully exploit future data.

Abstract

We present the strongest current cosmological upper limit on the sum of neutrino masses of < 0.18 (95% confidence). It is obtained by adding observations of the large-scale matter power spectrum from the WiggleZ Dark Energy Survey to observations of the cosmic microwave background data from the Planck surveyor, and measurements of the baryon acoustic oscillation scale. The limit is highly sensitive to the priors and assumptions about the neutrino scenario. We explore scenarios with neutrino masses close to the upper limit (degenerate masses), neutrino masses close to the lower limit where the hierarchy plays a role, and addition of massive or massless sterile species.

Figures (6)

• Figure 1: The current knowledge of neutrino masses and mixing between the interaction eigenstates as obtained from neutrino oscillation experiments Fukuda:1998Beringer:2012 for the three normal/active neutrinos ($\nu_e,\nu_\mu, \nu_\tau$). If the value of $\Delta m$ is large, the mass differences are much smaller than the neutrino masses, and the differences can be safely neglected. If $\Delta m$ is small, the ordering becomes important. Figure adapted from King:2013.
• Figure 2: $\Lambda\mathrm{CDM}$ fitted to Planck+WiggleZ as a function of $k_\mathrm{max}$. There is an excellent agreement between Planck and Planck+WiggleZ for all values of $k_\mathrm{max}$.
• Figure 3: WiggleZ power spectrum averaged (for visualisation only) over the seven survey regions and four redshift bins (black bars) shown with the best fit $\Lambda\mathrm{CDM}$ models for $k_\mathrm{max} = 0.2$$\,h\,\textrm{Mpc}^{-1}$ (red solid) and $k_\mathrm{max} = 0.3$$\,h\,\textrm{Mpc}^{-1}$ (blue solid) as well as the linear CLASS models for the same parameters (dotted, same colours). In the lower panel the models are compared after normalisation by the data values.
• Figure 4: One-dimensional parameter likelihoods for fitting $\Lambda\mathrm{CDM}_{3\nu}$ (left) and $\Lambda\mathrm{CDM}_{1+2\nu}$ (right) to various data combinations: Planck (dashed purple), Planck+BAO (dotted black), Planck+WiggleZ (dot-dashed green), Planck+BAO+WiggleZ (thick solid red), Planck+BAO+HST+WiggleZ (thin solid blue). The main effect of adding other observations to Planck is a tightening of the constraints on $\Omega_\mathrm{cdm}$, $H_0$ and $\sum m_\nu$ (top row). The improvement of adding WiggleZ is more significant for $\Lambda\mathrm{CDM}_{1+2\nu}$ than for $\Lambda\mathrm{CDM}_{3\nu}$ indicating that the fit is sensitive to the power spectrum shape.
• Figure 5: One-dimensional parameter probabilities comparing $\Lambda\mathrm{CDM}_{1+2\nu}$ (red), $\Lambda\mathrm{CDM}_{2+1\nu}$ (blue), and $\Lambda\mathrm{CDM}_{3\nu}$ (black) fits to BAO+Planck+WiggleZ (solid) and Planck+BAO (dashed). None of the preferred parameter shifts significantly between the different scenarios, only the $\sum m_\nu$ limit changes.