THE PRIMORDIAL HELIUM ABUNDANCE - SYSTEMATIC EFFECTS AND A NEW DETERMINATION

We present new high-quality spectrophotometric observations of 27 supe rgiant H II regions in 23 low-metallicity blue compact galaxies (BCGs) with oxygen abundance 12 + log (O/H) between 7.22 and 8.17 (Z./50 les s than or equal to Z less than or equal to Z./6), most of which were s elected from the first and second Byurakan objective-prism surveys. Co mbining this new data with our previous sample of 10 low-metallicity B CGs, we extract a subsample of 27 H II regions in 24 BCGs most appropr iate for the determination of the primordial helium abundance Y-p. We find that the most metal-deficient BCG known, I Zw 18 (Z similar to Z. /50), cannot be used for this purpose because of its abnormally low He I line intensities. We have examined critically the systematic effect s that may influence the determination of Y-p. We find that the effect s of the corrections for neutral helium and for underlying stellar abs orption in the He I lines, possible deviations from case B recombinati on theory, fluorescent enhancement of the He I line intensities, tempe rature fluctuations in H II regions, Wolf-Rayet stellar winds, and sup ernova shock waves to be small. The main effect comes from the particu lar set of atomic data used. The best set of atomic data for use in th e determination of Y-p is composed of Smits's He I emissivities and Ki ngdon & Ferland's collisional enhancement correction factors. This set gives the smallest dispersion of the data in the Y-O/H and Y-N/H plan es and corrects best the data for collisional enhancement effects. By extrapolating the Y versus O/H and Y versus N/H linear regressions to O/H = N/H = 0, we obtain Y,= 0.243 +/- 0.003 for both regressions, con siderably larger than the values derived before. In the framework of t he standard hot big bang nucleosynthesis model with a number of neutri no families N-v = 3 and a neutron half-lifetime tau(n) = 887 s, our ne w Y-p gives a baryon-to-photon number ratio eta = 3.5(-0.7)(1.0) x 10( -10), or a baryonic mass fraction Omega h(50)(2) = 0.05 +/- 0.01. Our Y-p determination is fully consistent with measurements of the primord ial Li-7 in Galactic halo stars and of (D + He-3) in the local interst ellar medium and the solar system, and consistent at the 2 sigma level with the primordial D abundance derived in a quasar absorption system by Tytler, Fan, & Buries. We derive a slope dY/dZ = 1.7 +/- 0.9, cons iderably smaller than the slopes obtained before and consistent with c hemical evolution models for star-forming dwarf galaxies with an outfl ow of well-mixed material. Extrapolation to solar metallicity with suc h a slope gives the correct solar helium mass fraction within the erro rs.