Franck Delahaye

The Solar Heavy-Element Abundances. I. Constraints from Stellar Interiors

The latest solar atmosphere models including non-LTE corrections and three-dimensional hydrodynamic convection simulations predict a significant reduction in the solar metal abundance. This leads to a serious conflict between helioseismic data and the predictions of solar interiors models. We demonstrate that the helioseismic constraints on the surface convection zone depth and helium abundance combined with stellar interiors models can be used to constrain chemical composition. A detailed examination of the errors in the theoretical models disfavors strongly (disagreeing at the 15 σ level with the seismic data) the proposed new low abundance, while the models constructed with the older and higher solar abundances are consistent (within 2 σ). We then use the sensitivity of the seismic properties to abundance changes to invert the problem and infer a seismic solar heavy-element abundance mix with two components: meteoritic abundances and the light metals CNONe. Seismic degeneracies between the best solutions for the elements arise for changes in the relative CNONe abundances and their effects are quantified. We obtain Fe/H = 7.50 ± 0.045 ± 0.003(CNNe) and O/H = 8.86 ± 0.041 ± 0.025(CNNe) on the logarithmic scale, where H = 12 for the relative CNNe mixtures in the Grevese & Sauval mixture; the second error term reflects the uncertainty in the overall abundance scale from errors in the C, N, and Ne abundances relative to oxygen. These are consistent within the errors with the previous standard solar mixture but in strong conflict with the low oxygen abundance inferred from the three-dimensional hydro models. Changes in the Ne abundance can mimic changes in oxygen for the purposes of scalar constraints. However, models constructed with low oxygen and high neon are inconsistent with the solar sound speed profile. Implications for the solar abundance scale are discussed.

X-ray absorption via Kα resonance complexes in oxygen ions

The Kα resonance complexes in oxygen ions O I‐O VI are theoretically computed, and resonance oscillator strengths and wavelengths are presented. The highly resolved photoionization cross-sections, with relativistic fine structure, are computed in the coupled channel approximation using the Breit‐Pauli R-matrix method. A number of strong Kα resonances are found to be appreciable, with resonance oscillator strengths f r > 0.1. The Kα resonance wavelengths of O I‐O VI lie in the relatively narrow wavelength range 22‐23.5 ˚ A, and the X-ray opacity in this region should therefore be significantly affected by K → L transitions in oxygen. The results should be useful in the interpretation of soft X-ray spectra observed from Chandra and

Atomic data from the Iron Project - XLIII. Transition probabilities for Fe V

An extensive set of dipole-allowed, intercom- bination, and forbidden transition probabilities for Fe V is presented. The Breit-Pauli R-matrix (BPRM) method is used to calculate 1:46 10 6 oscillator strengths for the allowed and intercombination E1 transitions among 3865 ne-structure levels dominated by conguration complexes with n 10 and l 9. These data are comple- mented by an atomic structure conguration interaction (CI) calculation using the SUPERSTRUCTURE program for 362 relativistic quadrupole (E2) and magnetic dipole (M1) transitions among 65 low-lying levels dominated by the 3d 4 and 3d 3 4s congurations. Procedures have been developed for the identication of the large number of ne-structure levels and transitions obtained through the BPRM calculations. The target ion Fe VI is represented by an eigenfunction expansion of 19 ne-structure levels of 3d 3 and a set of correlation congurations. Fe V bound levels are obtained with angular and spin symmetries SL and J of the (e + Fe VI) system such that 2S +1 =5 , 3, 1,L 10, J 8 of even and odd parities. The completeness of the calculated dataset is veried in terms of all possible bound levels belonging to relevant LS terms and transitions in correspondence with the LS terms. The ne-structure averaged relativistic values are compared with previous Opacity Project LS coupling data and other works. The 362 forbidden transition probabilities considerably extend the available data for the E2 and M1 transtions, and are in good agreement with those computed by Garstang for the 3d 4 transitions.

Comparison of Radiative Accelerations Obtained with Atomic Data from OP and OPAL

Microscopic diffusion processes (such as radiative levitation, gravitational settling, and thermal diffusion) in the outer layers of stars are important because they may give rise to surface abundance anomalies. Here we compare radiative accelerations (grad) derived from the new Opacity Project (OP) data with those computed from OPAL and some previous data from OP. For the case in which we have full data from OPAL (carbon, five points in the ρ-T plane), the differences in the Rosseland mean opacities between OPAL and the new OP data are within 12% and are less than 30% between new OP and previous OP data (OP1). The radiative accelerations grad differ at up to the 17% level when compared to OPAL and up to the 38% level when compared to OP1. The comparison with OP1 on a larger ρ-T space gives a difference of up to 40% for grad(C) and increases for heavier elements, reaching 60% for Si and 65% for S and Fe. We also constructed four representative stellar models in order to compare the new OP accelerations with prior published results that used OPAL data. The Rosseland means overall agree better than 10% for all our cases. For the accelerations, the comparisons with published values yield larger differences in general. The published OPAL accelerations for carbon are even larger relative to OP compared to those that our direct comparisons indicate. Potential reasons for this puzzling behavior are discussed. In light of the significant differences in the inferred acceleration rates, theoretical errors should be taken into account when comparing models with observations. The implications for stellar evolution are briefly discussed. The sensitivity of grad to the atomic physics may provide a useful test of different opacity sources.

Electron impact excitation of helium-like oxygen up to n = 4 levels including radiation damping

The primary x-ray diagnostic lines in He-like ions are mainly excited by electron impact from the ground level to the n = 2 levels, but at high temperatures n≥2 levels are also excited. In order to describe the atomic processes more completely, collision strengths are computed for O VII including for the first time all of the following: (i) relativistic fine structure, (ii) levels up to n = 4, and (iii) radiation damping of autoionizing resonances. The calculations are carried out using the Breit–Pauli R-matrix method with a 31-level eigenfunction expansion. Resonance structures in collision strengths are delineated in detail up to the n = 4 thresholds. For highly charged He-like ions radiation damping of autoionizing resonances is known to be significant in certain energy ranges. We investigate this effect in detail and find that while resonances are discernibly damped radiatively as the series limit n → ∞ is approached from below, the overall effect on effective cross sections and rate coefficients is found to be very small. Collision strengths for the principal lines important in x-ray plasma diagnostics, w, x, y and z, corresponding to the four transitions to the ground level 1s2(1S 0) ← 1s2p(1P 1o), 1s2p(3P 2o), 1s2p(3P 1o), 1s2s(3S 1), are explicitly shown. Significant differences are found with previous works for several transitions. The contribution from the resonances converging to the levels from the complex n = 4 is found to be significant for some transitions and can increase the rate by a factor of four. This work is carried out as part of the Iron Project-RmaX Network.

Resolution and accuracy of resonances in R-matrix cross sections

We investigate the effect of resonances in photoionization and recombination cross sections computed using the R-matrix method. Photoionization and recombination rates derived from high-resolution cross sections for oxygen ions are compared with earlier works with less resolution and accuracy, such as in the widely used Opacity Project data. We find significant differences in photoionization rates for O II metastable states, averaged over Planck functions corresponding to ionizing radiation fields, with respect to the intrinsic accuracy of the calculations and improved resolution. Furthermore, for highly charged ions other physical effects are also important. Recombination rate coefficients, averaged over a Maxwellian distribution, are extremely sensitive to the position and resolution of near-threshold resonances and radiation damping, in (e + O VII) ⟷ O VI + hν. Surprisingly, however, the effect on the monochromatic and the mean Rosseland and Planck bound-free opacities is relatively small, but may be potentially significant.