Crystallization experiments have been conducted on compositions along tholeiitic liquid lines of descent to define the compositional space for the development of silicate liquid immiscibility. Starting materials have 46-56 wt% SiO 2 , 11.7-17.7 wt% FeO tot , and Mgnumber between 0.29 and 0.36. These melts fall on the basaltic trends relevant for Mull, Iceland, Snake River Plain lavas and for the Sept Iles layered intrusion, where large-scale liquid immiscibility has been recognized. At one atmosphere under anhydrous conditions, immiscibility develops below 1,000-1,020°C in all of these compositionally diverse lavas. Extreme iron enrichment is not necessary; immiscibility also develops during iron depletion and silica enrichment. Variations in melt composition control the development of silicate liquid immiscibility along the tholeiitic trend. Elevation of Na 2 O ? K 2 O ? P 2 O 5 ? TiO 2 promotes the development of two immiscible liquids. Increasing melt CaO and Al 2 O 3 stabilizes a singleliquid field. New data and published phase equilibria show that anhydrous, low-pressure fractional crystallization is the most favorable condition for unmixing during differentiation. Pressure inhibits immiscibility because it expands the stability field of high-Ca clinopyroxene, which reduces the proportion of plagioclase in the crystallizing assemblage, thus enhancing early iron depletion. Magma mixing between primitive basalt and Fe-Ti-P-rich ferrobasalts can serve to elevate phosphorous and alkali contents and thereby promote unmixing. Water might decrease the temperature and size of the two-liquid field, potentially shifting the binodal (solvus) below the liquidus, leading the system to evolve as a single-melt phase.

Chemical data from the MESSENGER spacecraft revealed that surface rocks on Mercury are unusually enriched in sulfur compared to samples from other terrestrial planets. In order to understand the speciation and distribution of sulfur on Mercury, we performed high temperature (1200-1750 • C), lowto high-pressure (1 bar to 4 GPa) experiments on compositions representative of Mercurian lavas and on the silicate composition of an enstatite chondrite. We equilibrated silicate melts with sulfide and metallic melts under highly reducing conditions (IW-1.5 to IW-9.4; IW = iron-wüstite oxygen fugacity buffer). Under these oxygen fugacity conditions, sulfur dissolves in the silicate melt as S 2− and forms complexes with Fe 2+ , Mg 2+ and Ca 2+ . The sulfur concentration in silicate melts at sulfide saturation (SCSS) increases with increasing reducing conditions (from <1 wt.% S at IW-2 to >10 wt.% S at IW-8) and with increasing temperature. Metallic melts have a low sulfur content which decreases from 3 wt.% at IW-2 to 0 wt.% at IW-9. We developed an empirical parameterization to predict SCSS in Mercurian magmas as a function of oxygen fugacity ( f O 2 ), temperature, pressure and silicate melt composition. SCSS being not strictly a redox reaction, our expression is fully valid for magmatic systems containing a metal phase. Using physical constraints of the Mercurian mantle and magmas as well as our experimental results, we suggest that basalts on Mercury were free of sulfide globules when they erupted. The high sulfur contents revealed by MESSENGER result from the high sulfur solubility in silicate melt at reducing conditions. We make the realistic assumption that the oxygen fugacity of mantle rocks was set during equilibration of the magma ocean with the core and/or that the mantle contains a minor metal phase and combine our parameterization of SCSS with chemical data from MESSENGER to constrain the oxygen fugacity of Mercury's interior to IW-5.4 ± 0.4. We also calculate that the mantle of Mercury contains 7-11 wt.% S and that the metallic core of the planet has little sulfur (<1.5 wt.% S). The external part of the Mercurian core is likely to be made up of a thin (<90 km) FeS layer.

We present crystallization experiments on silicate melt compositions related to the lunar magma ocean (LMO) and its evolution with cooling. Our approach aims at constraining the primordial internal differentiation of the Moon into mantle and crust. We used graphite capsules in piston cylinder (1.35-0.80 GPa) and internally-heated pressure vessels (<0.50 GPa), over 1580-1020°C, and produced melt compositions using a stepwise approach that reproduces fractional crystallization. Using our new experimental dataset, we define phase equilibria and equations predicting the saturation of liquidus phases, magma temperature, and crystal/melt partitioning for major elements relevant for the crystallization of the LMO. These empirical expressions are then used in a forward model that predicts the liquid line of descent and crystallization products of a 600 km-thick magma ocean. Our results show that the effects of changes in the bulk composition on the sequence of crystallization are minor. Our experiments also show the crystallization of a silica phase at ca. 1080°C and we suggest that this phase might have contributed to the building of the lower anorthositic crust. Calculation of crustal thickness clearly shows that a thin crust similar to that revealed by GRAIL cannot have been generated through solidification of whole Moon magma ocean. We discuss the role of magma ocean depth, trapped liquid fraction (with implication for the alumina budget in the mantle and the crust), and the efficiency of plagioclase flotation in producing the thin crust. We also constrain the potential range of pyroxene compositions that could be incorporated into the crust and show that delayed crustal building during ca. 4% LMO crystallization on the nearside of the Moon may explain the dichotomy for Mg-number. Finally, we show that the LMO can produce magnesian anorthosites during the first stages of plagioclase crystallization.