HIGH-TEMPERATURE STABILITY OF LAURITE AND Ru–Os–Ir ALLOY AND THEIR ROLE IN PGE FRACTIONATION IN MAFIC MAGMAS: ERRATUM

As a result of an error in the JANAF tables in the free energy of formation (Gf) of the gas species S2O, H2S2 and HS (V. Kress, pers. commun., 2001), calculated values of oxygen and sulfur fugacities, f(O2) and f(S2), respectively, presented in Table 1 of Brenan & Andrews (2001; hereafter BA) are incorrect. A revised version of the program COHSmix was kindly supplied by Victor Kress, and new values of f(O2) and f(S2) have been calculated, and provided in Table 1 of this erratum (note also that there was a misprint in the rates of gas flow for experiment Ru7, and the correct values are provided). Inasmuch as the error is related to the calculation of the proportions of S-bearing gas species, we have checked the accuracy of calculated f(S2) using the pyrrhotite sulfur barometer of Toulmin & Barton (1964) at 900°C and log f(S2) of –1, –2 and –3 . On the basis of this assessment, the revised values of f(S2) quoted in Table 1 are considered to be accurate to within 0.2 log units. The calculated f(O2) was previously checked using the stability of solid oxide buffers (nickel – nickel oxide, iron – wüstite), and in this recent evaluation, we have also employed the NiO–Pd redox sensor (Pownceby & O’Neill 1994). This latter technique provided the most precise estimate of f(O2), and reproduced the COHSmix calculation to within 0.1 log units. As shown in Table 1, most values of f(O2) are unchanged relative to those we originally published, whereas the new values of f(S2) are shifted by as much as 0.7 log units. Although the implications of our results remain essentially unchanged, some specific interpretations of the data need to be modified in light of this correction. First, the maximum thermal stability of laurite estimated from experiments containing either Ru (1275 ± 25°C), or Ru plus other PGE (1265 ± 15°C), now applies to log f(S2) values of –1.5 and –1, respectively. Second, the intrinsic effect of temperature on coexisting laurite–alloy compositions (Fig. 8 of BA) cannot be defined by our results, as runs at 1250°C are at higher f(S2) than those at 1200°C. Judging from Figure 1 of BA, the effect of increased T is expected to be offset by higher f(S2), which probably accounts for the similarity in laurite–alloy pairs for those experiments. Lastly, experiments used to define the effect of f(S2) on the composition of coexisting laurite–alloy pairs (Fig. 9 of BA) differ only slightly in f(S2), again probably accounting for the relatively small shift in the composition of laurite–alloy pairs. We are currently performing a series of experiments to document these effects in more detail.