“…The resultant contribution of secondary X‐ray fluorescence from Cu grids as outlined in Jennings et al (), ~70% for both metal and silicate phases, is substantially larger than that observed in the current work, at ~50% in both phases. Although a detailed exploration into where this ~20% difference arises is far outside the scope of the current work, it is likely the cumulative result of multiple contributing factors, such as: - The approach taken by Jennings, Wade, and Llovet () to investigate the potential effects of secondary X‐ray fluorescence from Cu grids, which is largely based on Monte Carlo simulations in which idealized compositions are used (pure Fe for metal, CFMAS for silicate), and therefore does not fully encapsulate the experimental compositions of Mahan, Siebert, Blanchard, Badro, et al ().
- The sample lamella simulated, at 20 x 12 x 3 μm, are less than half the size of those typically found in Mahan, Siebert, Blanchard, Badro, et al () (approximately 20 x 30 x 3 μm), which may focus and/or enhance the effects of secondary X‐ray fluorescence.
- The simulations of Jennings, Wade, and Llovet () differ from typical DAC experimental run products as they assume that all silicate has melted to become basaltic glass, which is not the case (see Figures and S1; Mahan, Siebert, Blanchard, Borensztajn, et al, ; Mahan, Siebert, Blanchard, Badro, et al, ), and thus these simulations also cannot account for the size, structure, and/or geometry of this reaction zone.
- The simulations of Jennings, Wade, and Llovet () furthermore do not account for the Pt welding, gaps between the metal and silicate phases (from differential thermal expansion/contraction), surface topography, TEM grid orientation relative to samples, or any other morphological features and variables that are typical of actual DAC experiments.
…”