We present strong bounds on the sum of three active neutrino masses ( \sum m _ { \nu } ) using selected cosmological datasets and priors in various cosmological models .
We use the following baseline datasets : Cosmic Microwave Background ( CMB ) temperature data from Planck 2015 , Baryon Acoustic Oscillations measurements from SDSS-III BOSS DR12 , the newly released Type Ia supernovae ( SNe Ia ) dataset from Pantheon Sample , and a prior on the optical depth to reionization from 2016 Planck Intermediate results .
We constrain cosmological parameters with these datasets with a Bayesian analysis in the background of \Lambda CDM model with 3 massive active neutrinos .
For this minimal \Lambda CDM + \sum m _ { \nu } model we find a upper bound of \sum m _ { \nu } < 0.152 eV at 95 \% C.L .
Adding the high- l polarization data from Planck strengthens this bound to \sum m _ { \nu } < 0.118 eV , which is very close to the minimum required mass of \sum m _ { \nu } \simeq 0.1 eV for inverted hierarchy .
This bound is reduced to \sum m _ { \nu } < 0.110 eV when we also vary r , the tensor to scalar ratio ( \Lambda CDM + r + \sum m _ { \nu } model ) , and add an additional dataset , BK14 , the latest data released from the Bicep-Keck collaboration ( which we add only when r is varied ) .
This bound is further reduced to \sum m _ { \nu } < 0.101 eV in a cosmology with non-phantom dynamical dark energy ( w _ { 0 } w _ { a } CDM + \sum m _ { \nu } model with w ( z ) \geq - 1 for all z ) .
Considering the w _ { 0 } w _ { a } CDM + r + \sum m _ { \nu } model and adding the BK14 data again , the bound can be even further reduced to \sum m _ { \nu } < 0.093 eV .
For the w _ { 0 } w _ { a } CDM + \sum m _ { \nu } model without any constraint on w ( z ) , the bounds however relax to \sum m _ { \nu } < 0.276 eV .
Adding a prior on the Hubble constant ( H _ { 0 } = 73.24 \pm 1.74 km/sec/Mpc ) from Hubble Space Telescope ( HST ) , the above mentioned bounds further improve to \sum m _ { \nu } < 0.117 eV , 0.091 eV , 0.085 eV , 0.082 eV , 0.078 eV and 0.247 eV respectively .
This substantial improvement is mostly driven by a more than 3 \sigma tension between Planck 2015 and HST measurements of H _ { 0 } and should be taken cautiously .