Understanding an origin of palladium in metal-poor stars based on the non-LTE analysis of Pd~I lines

TL;DR

This study develops a comprehensive non-LTE treatment for Pd I by constructing a 106-level Pd I model atom and applying it to 48 metal-poor stars, deriving Pd, Sr, Ba, and Eu abundances. The NLTE analysis weakens Pd I lines and yields positive abundance corrections, aligning solar Pd with meteoritic values and reducing LTE discrepancies between giants and dwarfs. The Pd–Eu correlation across −1.71 ≤ [Fe/H] ≤ −0.56 indicates roughly 70% r-process and 30% s-process contributions to Pd, with pure r-process Pd in two r-II stars and substantial r-process dominance in two very metal-poor stars; two Sr-enhanced VMP stars suggest additional Sr and Pd sources from fast-rotating massive stars. Overall, NLTE is crucial for constraining Galactic Pd evolution and identifying multiple nucleosynthetic channels contributing to Pd alongside Sr, Ba, and Eu.

Abstract

Palladium is one of poorly observed neutron-capture elements. Abundance determinations for stellar samples covering a broad metallicity range are needed for better understanding the mechanisms of Pd synthesis during the Galaxy evolution. We aim to obtain accurate abundances of Pd for the Sun and the sample of metal-poor stars based on the non-local thermodynamic equilibrium (non-LTE) line formation for Pd~I. We present a new, comprehensive model atom of Pd~I. Abundances of Pd, Sr, Ba, and Eu were derived for 48 stars from the non-LTE analyses of high resolution and high signal-to-noise ratio spectra provided by the ESO archives. Non-LTE leads to weakened Pd~I lines and positive non-LTE abundance corrections growing from 0.2~dex for the solar lines up to 0.8~dex for the lines in the most luminous star of the sample. Depending on a treatment of inelastic collisions with hydrogen atoms, the solar non-LTE abundance amounts to log eps = 1.61+-0.02 to 1.70+-0.02 and agrees within the error bars with the meteoritic abundance log eps_met = 1.65. Non-LTE largely removes the discrepancies in the LTE abundances between the giant and dwarf stars of similar metallicities. Palladium tightly correlates with Eu in the -1.71 < [Fe/H] < -0.56 range indicating the r- and s-process contributions to Pd synthesis of approximately 70% and 30%, respectively. Palladium is of pure r-process origin in our two r-II stars, and a dominant contribution of the r-process to the Pd abundances is found for another two very metal-poor (VMP, [Fe/H] < -2) stars. The two VMP stars, which are strongly enhanced with Sr relative to Ba and Eu, reveal also enhancements with Pd. We propose that the source of extra Sr and Pd in these stars are VMP, fast rotating massive stars. Non-LTE is essential for obtaining the observational constraints to future models of the Galactic Pd evolution.

Figures (8)

• Figure 1: Energy levels of Pdi, as presented in the model atom. The spectral lines used in Pd abundance analysis arise in the transitions shown as continuous lines. The term designations correspond to atomic structure calculations of Kurucz2017.
• Figure 2: Departure coefficients, b, for the selected levels of Pdi as a function of $\log\tau_{5000}$ in the model atmosphere 5780/4.44/0. The tick marks indicate the location of line center optical depth unity for the Pdi 3242 (1) and 3404 Å (2) lines.
• Figure 3: Best NLTE ($S_{\!\rm H}$ = 0.1) fits (continuous curves) of the Pdi lines in the solar disk-center intensity spectrum Delbouille1973. The obtained NLTE abundances are indicated in Table \ref{['tab:sun']}. In each panel, the dashed curve shows the LTE profile computed with the NLTE abundance derived from this line and the dotted curve is the synthetic spectrum without the contribution of the Pdi line.
• Figure 4: Best fit (continuous curve) of the 3515.8 -- 3518.5 Å range in the solar disk-center intensity spectrum Delbouille1973. The dotted curve is the synthetic spectrum without the contribution of the Pdi 3516 Å line. The dash-dotted line shows a local continuum level that was adopted for Pdi 3516 Å when deriving the element abundance listed in Table \ref{['tab:sun']}.
• Figure 5: NLTE (stars) and LTE (open circles) abundance ratios [Pd/Fe] for the sample stars. The smaller and bigger size symbols correspond to the dwarfs ( log g $> 3.0$) and the giants ( log g $\le 3.0$), respectively. For each star, the error bar corresponds to $\sigma_{\rm tot}$.