INFLUENCE OF FeO ON THE PARTITIONING OF NA BETWEEN OLIVINE AND SILICATE MELT – IMPLICATIONS FOR THE BEHAVIOR OF ALKALIS DURING CHONDRULE FORMATION

Introduction: It has been proposed that chondrules behaved as chemically closed system with respect to Na [1-5]. This implies that Na did not evaporate from the chondrule melt after onset of olivine crystallization. High temperatures, low pressure and low fO 2 during chondrule formation, however, should lead to rapid loss (minutes) of Na [e.g., 6]. This is not observed. Understanding the relation between Na in chondrule olivine and silicate melt requires knowledge of the olivine/melt partitioning coefficient D Na. [5] determined D Na to 0.0031±0.0007. [1], however, suggest much higher D Na for their Semarkona chondrules. In order to explain differences in D Na in terms of varying FeO contents, we have conducted experiments in FeO-free systems and have re-evaluated data from Semarkona chondrules. Methods: Various FeO-free starting materials (CMAS-Na) with 0 – 6.7 wt.% Na 2 O were enclosed in platinum capsules, shortly molten at 1600°C and heated for 3 – 4 days to 1300°C, 1350°C and 1450°C, respectively. Samples where quenched at the end of the experiments. After major element analyses, Na in olivine was analyzed by means of high current EMPA technique [5]. The detection limit for Na 2 O in olivine is at 15 µg/g. Results and Discussion: For FeO-free systems, we determined a partition coefficient of Na between olivine and melt of D Na = 0.00139±0.00039 (1σ). No temperature dependence within the experimental temperature range was found. The value is lower than the D Na that was determined by [5]. The difference may arise from different FeO contents with 7-13 wt.% FeO in the experiments by [5]. Our datum and the experimental results from [5] suggest that D Na (ol,melt) is a function of X fa with DNa = 0.00139 + 0.000186*FeO ol [wt.%]. The partitioning of Ca is similarly a function of FeO in olivine [7]. Plots of apparent D Na and corresponding FeO [1-5] from Semarkona type-I chondrule olivine (core/bulk, rim/meso) and type-II chondrule olivine (only core/bulk) can be well-explained in terms of D Na as function of FeO in olivine. Apparent D Na data from type-II chondrule olivine rims (FeO-rich) and corresponding mesostasis do not agree well with the proposed linear relation between D Na and FeO. Instead apparent D Na values are systematically lower as expected from D Na = f(FeO). The lower D Na may result from formation of the rims at low T (<1000°C, possibly large degree of undercooling) …