3.2. Helium-4

As the second most abundant nuclide in the Universe (after hydrogen), the abundance of ⁴He can be determined to high accuracy at sites throughout the Universe. However, as stars evolve they burn hydrogen to helium and the he in the debris of stellar evolution contaminates any primordial ⁴He. Since any attempt to correct for stellar evolution will be inherently uncertain, it is sensible to concentrate on the data from low-metallicity sites. Extragalactic regions of hot, ionized gas (H II regions) provide such sites, where the helium is revealed via emission lines formed when singly and doubly ionized helium recombines. As with deuterium, current data provide ambiguous estimates of the primordial helium abundance. Since the differences (Y = 0.010) are larger than the statistical uncertainties ( ± 0.003), systematic errors likely dominate. Among the currently most likely sources of such errors are uncertain corrections for collisional excitation in helium, uncertain corrections for unseen neutral helium and/or hydrogen, and underlying stellar absorption (leading to an underestimate of the true strength of the helium emission lines). In contrast, since the most metal-poor of the observed regions have metallicities of order 1/50 - 1/30 of solar, the extrapolation from the lowest metallicity regions to truly primordial introduces an uncertainty no larger than the statistical error.

Using published data ([14, 15]) for 40 low-metallicity regions, one group [13] finds: Y_P = 0.234 ± 0.003. In contrast, from an independent data set of 45 low-metallicity regions another group [16] infers Y_P = 0.244 ± 0.002. Clearly, these results are statistically inconsistent. It is crucial that high priority be assigned to further H II region observations to estimate/avoid the systematic errors. Until then, since the error budget for Y_P is likely dominated by systematic rather than statistical uncertainties, in what follows, a generous range for Y_P will be adopted: 0.228 Y_P 0.248.