Primordial helium has narrowed enough at the Large Binocular Telescope to weigh the neutrino families

Where the measurement now stands

The quantity is not new. Helium forged in the first five minutes of the universe has been read off pristine gas for decades. What has moved is the error bar. From 130 hours on the Large Binocular Telescope, across 15 of the most pristine small and remote galaxies, more than 10 helium lines and 15 hydrogen lines were measured together, and the uncertainty on the primordial helium abundance came down to 0.5 percent, three times better than previous standards.^[1]

The work that bought that factor of three was done inside the instrument. Archival standard-star spectra spanning four years were used to determine a wavelength-dependent uncertainty in the MODS spectral response, which improved the relative uncertainties between emission lines. With that in hand, individual H II regions can be delivered at roughly 2 percent or less. The MODS spectrographs themselves took 12 years to build from conception to first light on sky.^[1]

What the number weighs

On its own a helium mass fraction says little. Taken with standard Big Bang nucleosynthesis and the half-life of the neutron, it becomes a direct measurement of the number of neutrino families present when the universe was minutes old. That is why the precision matters: a determination of the primordial helium abundance at about 0.5 percent is enough to supply an independent constraint on the effective number of neutrino families of about 3 percent, and only at that level does the measurement begin to bear on physics beyond the Standard Model.^[1]

Where the remaining error lives is the interesting part. The project's own account puts the improvement on calibrating the MODS spectral response rather than on collecting more photons, so the present limit on this measurement looks instrumental. The prosaic rival reading deserves its full weight: even in the most metal-poor H II regions, a residual astrophysical correction for helium made by stars rather than in the first five minutes may dominate what is left, in which case the share remaining for further calibration work is small.^[1]

The next measurement

The LBT Yp Project has set out its sample, its observations and its methodology, and the independent constraint it aims at for the effective number of neutrino families is about 3 percent. If that constraint is published from the optimized low-metallicity sample, the observable test is narrow and checkable: whether it lands near 3 percent, and whether the three families of the Standard Model sit inside it.^[1]