Heavy Element Abundances at High RedshiftsWallace Sargent Palomar Observatory, California Institute of TechnologyAy 219; May 26, 2010Tuesday, May 25, 2010Outline•QSO absorption lines: introduction•Abundances in DLA systems•Very low metallicity DLA systems•C and O abundances in the IGM at z ~ 2.5•Abundances at z ~ 6: •C, O and Si and Very Massive Stars •Fe/O and the origin of Fe at early timesTuesday, May 25, 2010Galaxies at Low Redshifts•Only Dwarf galaxies are metal-poor overall•Dwarf spheroidals have [Fe/H] ~ -2•Dwarf irregulars (BCGs) have O/H ~ 1/50 solar (and no lower)•None of these systems appear to be very young•We have not found local gas clouds devoid of stars or heavy elements (?)3Tuesday, May 25, 2010Tuesday, May 25, 2010Observed frameTuesday, May 25, 2010The Cosmic web at z ~4 25MpcTuesday, May 25, 2010Abundances as a function of H I column densityCorrected for dust depletion[Z/H]H I atoms cm−2Tuesday, May 25, 2010Abundances at high redshifts - questions:•When were the first heavy elements produced ?•Initial mass function of the ‘First Stars’?•How has the mass in elements heavier than H evolved in time?•Distribution of [Z/H] in the IGM?•Are there regions (e.g. in the voids) devoid of heavy elements?Tuesday, May 25, 2010Lyman-alpha forest gets thicker with increasing redshiftUniverse gets denser with increasing z and density of ionizing sources decreases above z ~2.5nHI∝ n2H/nγnH∝ (1 + z)3Tuesday, May 25, 2010Tuesday, May 25, 2010Evidence for dust in DLA absorption systems. Zn is not depleted in the Galactic ISM -- low condensation temperature.Zn/Fe high in gas phase because Fe depleted into dust grainsTuesday, May 25, 2010Tuesday, May 25, 2010Tuesday, May 25, 2010Si/Fe ratios in DLA systems. Fe is dust depleted at high Si/H. At low Si/H values the Si/Fe ratio is typical of Type II SNe.Tuesday, May 25, 2010Tuesday, May 25, 2010Tuesday, May 25, 2010Tuesday, May 25, 2010New results showing O/H for sampleTuesday, May 25, 2010Tuesday, May 25, 2010Search for very metal poor DLA absorbers in Prochaska SDSS sample: conclusions:-•No absorbers with M/H convincingly below 0.001 solar - there is a ‘floor’•Usually one velocity component - most metal poor DLAs are dwarf systems as in local universe•C/O rises probably to super-solar value for [O/H] ~ -3 continuing a trend observed in DLAs and halo stars at lower [O/H]Tuesday, May 25, 2010Abundances in the IGM at High Redshifts•Almost all of the baryons are in the IGM.•The IGM is highly ionized-- log(N(H I)/N(H)) ~ -5.•General IGM ionized mostly by observed QSOs at z ~ 2- 5.•Abundant ions: C IV, Si IV, N V, O VI, C II, Si II.•Relative strengths depend on IGM density (big contrasts between the voids and the walls and filaments).•Few elements can be observed.Tuesday, May 25, 2010Element Abundance Distributions in the IGM•Best ions are C IV 1548, 1550A and O VI 1032, 1036A doublets•C IV and O VI constitute roughly 20 percent of the C and O in the tenuous IGM•O VI is always in the Lyman-alpha forest•At z ~ 2.5 have thin forest and bright QSOs to use as background sources for spectroscopyTuesday, May 25, 2010Tuesday, May 25, 2010Predicted ionization fractions for O and C in the IGM as a function of H I column density. O VI predominates over C IV for log N(H I) < 15. Simcoe, Sargent and Rauch (2004)Tuesday, May 25, 2010Contribution of different H I column densities to and their mean dependance on overdensity. The maximum contribution is from lines with . Our O VI survey is complete for overdensities and for column densities . Ωbaryon1013≤ N (HI) ≤ 1014N (HI) ≥ 3 × 1013ρ/¯ρ ≥ 2 Simcoe, Sargent and Rauch (2004)Tuesday, May 25, 2010Haardt-Madau metagalactic ionizing spectrumergs cm−2s−1Hz−1ster−1``Proximity effect’’Tuesday, May 25, 2010Tuesday, May 25, 2010Expected strength of O VI 1032 for [O/H] = -2.5 Tuesday, May 25, 2010Plot of N(O VI) versus N(H I). The black points are upper limits. The curve shows the expected relation for [O/H] = -2.5Simcoe, Sargent and Rauch (2004)O becomes O VO becomes O VIITuesday, May 25, 2010Tuesday, May 25, 2010The distribution of [O/H] and [C/H] for absorption systems at z ~ 2.5. Simcoe, Sargent and Rauch (2004)Tuesday, May 25, 2010Cumulative fraction of H I lines (with log N(H I) > 13.3) as a function of [O/H] and [C/H] Simcoe, Sargent and Rauch (2004)Tuesday, May 25, 2010Cumulative fraction of baryons by mass as a function of [O/H] and [C/H]. About 30 percent of the mass in the IGM has [Z/H] < -3. Simcoe, Sargent and Rauch (2004)Tuesday, May 25, 2010Tuesday, May 25, 2010C IV, O I and the end of reionization?•Below z ~ 5 IGM is highly ionized•O VI and C VI predominate•What happens at z ~ 6?•C IV 1548, 1550 redshifted to 10,848, out of the HIRES regime and into NIRSPEC echelle (R ~ 15,000) territory•O I 1302, C II 1335, Si II 1304 move to ~9240, still observable with HIRES Tuesday, May 25, 2010Lyman-alpha forest gets thicker with increasing redshiftUniverse gets denser with increasing z and density of ionizing sources decreases above z ~2.5nHI∝ n2H/nγnH∝ (1 + z)3Tuesday, May 25, 2010Spectrum (without axes) of SDSS J0002+2550 (z = 5.93)Lyman-alpha forest:•Traces bulk of the IGM•Saturates at z ~ 6Metal lines-Trace dense regions- Still useful at z > 6Significance of z ~ 6Tuesday, May 25, 2010Observed wavelengths at z ~ 6Ly alpha 1216 8512O I 1302 9114C II 1334 9338Si II 1526 10682 C IV 1548,50 10843Fe II 2344 16408Rest (A) Observed (A)HIRES NIRSPEC(R = 45000) (R = 25000)Tuesday, May 25, 2010Examples of NIRSPEC echelle data on z ~ 6 QSOs(Typical exposure times 10-12 hours: R = 13,000)Mg II 2797,2803; z ~ 2.77 1 sigma noiseNo C IVFe II 2374 z = 3.404Tuesday, May 25, 2010Detectability of C IV in Keck NIRSPEC data(Typical exposure times 10-12 hours)Tuesday, May 25, 2010Tuesday, May 25, 2010Contribution of C IV to Omega as a function of zRyan-Weber et al. (2009)Becker et al. (2009)Tuesday, May 25, 2010C IITuesday, May 25, 2010The HIRES z > 5 sample•12 Keck nights•Upgraded detector•9 QSOs at z > 4.8•3 at z > 6•Sensitive to O I 1302 over a narrow redshift range:•Max