Mineral-melt calcium isotope fractionation factors constrained using ab initio molecular dynamics simulations and their implications to calcium isotope effects during partial melting in the upper mantle
Description
This data set contains all data tables from "Mineral-melt calcium isotope fractionation factors constrained using ab initio molecular dynamics simulations and their implications to calcium isotope effects during partial melting in the upper mantle" by Yonghui Li, Justin Hardin, Wenzhong Wang, Zhongqing Wu and Shichun Huang, GCA
Files
Steps to reproduce
Our approach is to use ab initio molecular dynamic (AIMD) simulations to compute the Ca isotope β values (reduced partition function ratio, RPFR) of both melts and minerals. The β value of a given phase measures the equilibrium isotope fractionation factor between that phase and an ideal gas of Ca atoms, and it is linearly correlated with 1/T2 where T is temperature in Kelvin. Then, the mineral-melt Ca isotope fractionation factor is calculated as the difference between the β value of the mineral and that of the melt. We note that β values computed using different approaches and different software may have large systematic differences (e.g., Wang et al. (2017b); Li et al. (2022), and see Section 3 for a detailed discussion), and mineral-melt Ca isotope fractionation factors should not be calculated using β values obtained from different approaches. In this paper, we will use melt and mineral β values computed using the same approach. Our goal is to use the Δ44/40Camineral-melt from our ab initio calculations to update the calculations exploring the Ca isotope effects during partial melting of the mantle (e.g., Zhang et al., 2018; Chen et al., 2020b; Soderman et al., 2022; Antonelli et al., 2023c). Specifically, two approaches are used in this study. To model the Ca isotope effects during partial melting of spinel peridotites at 1 and 2 GPa, we follow the approach of Zhang et al. (2018) to use the pMELTS program (Ghiorso et al., 2002) to simulate partial melting of the mantle (Table 5). Specifically, in the batch melting model, the starting composition is re-equilibrated at each model temperature and pressure condition using pMELTS at a temperature step of 5-10 °C. In the fractional melting model, the residual composition from the previous step is used as the new starting composition of the following step. In the dynamic melting model, the calculation is completed the same way as in the fractional melting model, but with a porosity of 5%. To model the Ca isotope effects during partial melting with garnet as a residual phase (≥ 3 GPa), we use phase relationships and compositions from high P-T partial melting experiments (Walter, 1998; Yaxley and Green, 1998; Pertermann and Hirschmann, 2003; Dasgupta et al., 2007; Gerbode and Dasgupta, 2010; Thomson et al., 2016). Ca isotope effects in melts and residues are calculated using Δ44/40Camineral-melt (Equations 9-13), phase relationships, and mineral and melt compositions, with the assumption of all sources having a BSE-like 44/40CaSRM915a (0.96) (Kang et al. 2017; Antonelli and Simon, 2020). The results are summarized in Figures 8-10 and Tables 5-6. More detailed information is given in Supplementary Table S2, including phase proportions and their compositions under each pressure and temperature condition, as well as the Ca fraction in each mineral and melt phase. In this section, we will compare our simulation results to measurements of natural rocks and discuss their similarities and differences.