The primordial deuterium abundance of the most metal-poor damped Lyman-alpha system

TL;DR

This work reports a high-precision measurement of the primordial deuterium abundance from the most metal-poor DLA to date, obtaining a weighted-mean $oxed{\log_{10}(D/H)_{\rm P} = -4.5940 \pm 0.0056}$ (or $10^{5}(D/H)_{\rm P} = 2.547 \pm 0.033$). Using Monte Carlo BBN with updated cross sections, including the recent $d(p,\gamma)^3{\rm He}$ rate, yields $100\Omega_{\rm B,0}h^{2}_{\rm (BBN)} = 2.156 \pm 0.017 \pm 0.011$, which is $\sim2.3\sigma$ below the Planck CMB value; replacing the rate with an empirical determination gives $100\Omega_{\rm B,0}h^{2}_{\rm (BBN)} = 2.260 \pm 0.034$, in closer agreement. The paper also reports tentative evidence for a mild D/H decline with metallicity and discusses future prospects, including LUNA measurements and the potential for hundreds of precise D/H measurements with upcoming 30–40 m class telescopes. Overall, the results provide a stringent cosmological constraint on baryon density and highlight a small, ongoing tension between BBN and CMB results depending on nuclear inputs.

Abstract

We report the discovery and analysis of the most metal-poor damped Lyman-alpha (DLA) system currently known, which also displays the Lyman series absorption lines of neutral deuterium. The average [O/H] abundance of this system is [O/H] = -2.804 +/- 0.015, which includes an absorption component with [O/H] = -3.07 +/- 0.03. Despite the unfortunate blending of many weak D I absorption lines, we report a precise measurement of the deuterium abundance of this system. Using the six highest quality and self-consistently analyzed measures of D/H in DLAs, we report tentative evidence for a subtle decrease of D/H with increasing metallicity. This trend must be confirmed with future high precision D/H measurements spanning a range of metallicity. A weighted mean of these six independent measures provides our best estimate of the primordial abundance of deuterium, 10^5 (D/H)_P = 2.547 +/- 0.033 (log_10 (D/H)_P = -4.5940 +/- 0.0056). We perform a series of detailed Monte Carlo calculations of Big Bang nucleosynthesis (BBN) that incorporate the latest determinations of several key nuclear cross sections, and propagate their associated uncertainty. Combining our measurement of (D/H)_P with these BBN calculations yields an estimate of the cosmic baryon density, 100 Omega_B,0 h^2(BBN) = 2.156 +/- 0.020, if we adopt the most recent theoretical determination of the d(p,gamma)3He reaction rate. This measure of Omega_B,0 h^2 differs by ~2.3 sigma from the Standard Model value estimated from the Planck observations of the cosmic microwave background. Using instead a d(p,gamma)3He reaction rate that is based on the best available experimental cross section data, we estimate 100 Omega_B,0 h^2(BBN) = 2.260 +/- 0.034, which is in somewhat better agreement with the Planck value. Forthcoming measurements of the crucial d(p,gamma)3He cross section may shed further light on this discrepancy.

Figures (7)

• Figure 1: Top panels: The flux calibrated Hi Ly$\alpha$ absorption profile (black histogram) is shown for the DLA at $z_{\rm abs}=2.853054$ towards the quasar J1358$+$0349. The best-fitting quasar continuum model (blue long-dashed curves) and the best-fitting absorption profile (red line) are overlaid. The green dashed line indicates the fitted zero-level of the data. The spectrograph used to take the data is indicated in the upper left corner of each panel. Bottom panels: Same as the top panels, but with the quasar continuum normalized, and the data are plotted in the rest-frame of the DLA. The absorption feature that is fit near a rest wavelength of 1206.5 is a combination of the Si iii absorption from the DLA and an unrelated blend.
• Figure 2: A selection of the metal absorption lines associated with the DLA at $z_{\rm abs}=2.853054$ towards J1358$+$0349 that are used in our analysis. The best-fitting model (red line) is derived from a simultaneous fit to both the UVES and HIRES data. However, in these panels we only show the data (black histogram) and corresponding model for the dataset with the higher S/N near the absorption line. In all panels, the best-fitting zero-level of the data (short green dashed line) has been removed, and the continuum has been normalized (long blue dashed line). Note that we have used a different y-axis scale for the top row of panels to emphasize the weakest absorption features. The red tick marks above the spectrum correspond to the locations of the absorption components of the annotated ion (see Table \ref{['tab:compstruct']}). The green tick marks in the N iii$\lambda989$ panel are for a blend with Si ii$\lambda989$, the latter of which is largely determined from the multitude of other Si ii absorption lines. The absorption at $-25~{\rm km~s}^{-1}$ in the N ii$\lambda1083$ panel is assumed to be an unrelated blend.
• Figure 3: The black histogram shows our HIRES data (left panels) and UVES data (right panels), covering the Hi and Di Lyman series absorption lines from Ly$\alpha$--Ly7 (top to bottom panels, respectively). Our best-fitting model is overlaid with the solid red line. The plotted data have been corrected for the best-fitting zero-level (short green dashed line), and are normalized by the best-fitting continuum model (long blue dashed line). Tick marks above the spectrum indicate the absorption components for Hi (red ticks), and Di (green ticks).
• Figure 4: Same as Fig. \ref{['fig:lyseriesa']}, for the Hi and Di transitions Ly8--Ly15. Note that the leftmost set of red tick marks in the bottom panels indicate the Hi Ly15 absorption components, while the central red tick marks in these panels indicate Hi Ly14 absorption.
• Figure 5: We plot the current sample of high quality primordial D/H abundance measurements (symbols with error bars) as a function of the oxygen abundance. The green symbol (with the lowest value of [O/H]) corresponds to the new measurement reported here, and the blue symbols are taken from Coo14. The red dashed and dotted horizontal lines indicate the 68 and 95 per cent confidence interval on the weighted mean value of the six high precision D/H measures listed in Table \ref{['tab:dhmeasures']}. The right axes show the conversion between D/H and $\Omega_{\rm B,0}\,h^{2}$ for the Standard Model. The conversion shown in the left panel uses the recent theoretical determination of the $d(p, \gamma)^3$He reaction rate (and its error) by Mar16, while the right panel uses an empirical $d(p, \gamma)^3$He rate and error based on the best available experimental data (see NolBur00 and NolHol11 for a critical assessment of the available experimental data). In both panels, the gray horizontal band shows the Standard Model D/H abundance based on our BBN calculations (see text) and the universal baryon density determined from the CMB temperature fluctuations Efs15. The dark and light shades of gray represent the 68 and 95 per cent confidence bounds, respectively, including the uncertainty in the conversion of $\Omega_{\rm B,0}\,h^{2}$ to D/H (0.83 per cent for the left panel and 2.0 per cent for the right panel). The Standard Model value displayed in the left panel is 0.02 dex lower in $\log_{10}$(D/H) than that shown in Figure 5 of Coo14. This shift is largely due to the updated Planck results Efs15, and the updated theoretical $d(p,\gamma)^{3}{\rm He}$ reaction rate Mar16.