c

UCRL-JC- 123237 PREPRINT

c ~ p - q 4 0 3 i ~ 6 - - - J

Metal-Sulfide Melt Non-Interconnectivity in Silicates, Even at High Pressure, High Temperature,

and High Melt Fractions

W.G. Minarik F.J. Ryerson

This paper was prepared for submittal to the 27th Annual Lunar & Planetary Science Conference

Houston, TX March 18-22, 1996

January 1996 I \

This is a prepnnt of a paper intended for publication in a journal or proceedings Since changes may be made before publication, this preprint is made available with the understanding that it will not be cited or reproduced without the permission of the author.

DISCLAIMER

This document was prepared as an account of work sponsored by an agency of the United States Government. Neither the United States Government nor the University of California nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its tlse would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise, does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or the University of California. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or the University of California, and shall not be used for advertising or product endorsement purposes.

METAL-SULFIDE MELT NON-INTERCONNECTIVITY IN SILICATES, EVEN AT HIGH PRESSURE, HIGH TEMPERATURE, AND HIGH MELT FRACTIONS. William G. Minarik and Frederick J. Ryerson, Earth Sciences, L-202, Lawrence Livermore National Laboratory, Livermore, CA 9455 1 ; e-mail: minarik@llnl.gov

We have investigated the textural microstructure of a iron-nickel-sulfur melts in contact with olivine, pyroxene, and the modified-spinel polymorph of olivine. The experiments were conducted at 1500°C and pressures ranging from 1 to 17 GPa. For compositions more metal- rich than the monosulfide, including the eutectic composition, the metal sulfide melt has a dihedral angle greater than 60" and does not form an interconnected grain-edge fluid. Increasing pressure does not measurably alter the dihedral angles. Textural evolution results in coarsening of the sulfide melt pockets, resulting in large pockets surrounded by many silicate grains and separated from one another by melt-free grain edges. Chemical communication between these large pockets is limited to lattice and grain-boundary diffusion.

Due to the large interfacial energy between sulfide melt and silicates, sulfide melts are unable to separate from solid silicate via grain-boundary percolation and remain stranded in isolated melt pockets. Sulfide melt in excess of the critical melt fraction (5-25%) [ 1,2] will develop a transient interconnectivity as sulfide collects into larger melt pockets and interconnectivity is pinched off. Efficient separation of core-forming sulfide melts from silicate requires either melting of the silicate matrix or a very large fraction of metal-sulfide melt (perhaps as large as 40% [3]).

Experimental: Fe-Ni-S of the Urakawa et al. [4] 10 GPa composition was mixed with olivine or pyroxene

powders from a San Carlos xenolith ground to c approx. 50 pn size. Oxide was added to insure oxygen saturation and to help buffer silica activity, and extra iron was added to replace iron lost due to exchange with the silicates. The mixed powder was run in graphite capsules (diamond in the high-P experiments) in either piston-cylinder (e 4 GPa) or multi-anvil (9-17 GPa) apparatus. A time series shows that four hours is sufficient for dihedral angles to reach their equilibrium values, although grain growth and coarsening continues for days. Results are being obtained for other high-P phases in addition to beta-spinel.

Each melt-silicate contact angle on polished sections was either digitized optically with a through-the-scope LED display or from scanned backscattered electron (BSE) images. All measurable angles within an area of the section were digitized to avoid adding selection bias. The measured angles are spread through a distribution because any particular observed angle is a slice through the three-dimensional grain edge at a random orientation. The median angle of the distribution is an approximation of the true dihedral angle in systems with a single characteristic wetting angle [5 ] . A dihedral angle of 60" separates melts that wet the grain edges and form interconhected melt channels ( 4 0 " ) from melts that do not wet grain edges and remain in pockets at grain corners isolated by dry grain edges.

Results and Interpretation:

70" to lOO", depending mostly on sulfur content in the melt and only weakly as a function of silicate phase or pressure. During grain coarsening, sulfide melt pockets also coalesce into larger pockets surrounded by melt-free grain boundaries, thereby reducing the area of surface in contact with the silicate and lowering the overall interfacial energy of the system. As the pockets increase in size, the distance between them increases, slowing their further growth [6]. These results are consistent with those reported by Jurewicz and Jones [2, 71 and recently by

Photos of several experiments are appended as figures 1 and 2. Dihedral angles range from

SULFIDE MELT/SILICATE TEXTURE: W. G. Minarik and F. J. Ryerson

Shannon and Agee [8] for sulfide melt in contact with multiple silicate phases formed from melting an ordinary chondrite. These results conflict with the low angles reported by Herpfer and Larimer [9], although their melt compositions may be more sulfur-rich than reported [ 101.

Our study supports the hypothesis that metal-rich melts can not separate from silicates by porous flow at low melt fraction in bodies less than 3000 km in diameter. The pressure at the center of the Earth’s moon is 4.6 GPa and the Moon is likely depleted in sulfur. Hence, separation of a sulfur-poor metal core in the Moon would require extensive silicate melting. The pressures of this study correspond to the upper 400 km of the present Earth’s mantle. Late- stage accretion of metallic material with energies insufficient to melt the mantle to below this depth can trap metallic melts in the upper mantle. A pond of metallic melt at the bottom of a impact-heated zone would only be able to join the core if it was massive enough to sink diapirically through the solid silicate below; it would not be able leak downward by percolating along olivine grain boundaries. The Earth’s core may have been assembled from the cores of accreted planetesimals whose compositions reflect high-temperature, low pressure equilibrium, and then been modified over time by interaction with mantle silicates at core-mantle boundary pressures. Non-wetting metal-sulfide melts provide a physical mechanism for models of core formation which invoke incomplete separation of metal sulfide core from silicate mantle.

1) A reflected light photo of sulfide melt within an olivine and orthopyroxene matrix, contained within a graphite capsule (1 GPa, 1500OC, 2 days). The capsule is approx. 2mm across. The 8.5 vol.% sulfide has shown no tendency to segregate from the silicate matrix. 2) A BSE photo of the same experiment, The bright sulfide melt (quenched to a mixture of metal-rich and sulfur- rich phases) is excluded from grain boundaries and restricted to grain corners. Several sulfide pockets have coarsened and are enclosed by more than four silicate grains.

[l] von Bargen N. and Waff H. S . (1986) JGRQ1, 9261. [2] Jurewicz S . R. and Jones J. H. (1995) LPSC XXVI,2,709 [3] Taylor G. J. (1992) JGR,97, 14717. [4] Urakawa S . et al. (1987) in High Pressure Research in Mineral Phvsics, 95. [5] Jurewicz S . R. and Jurewicz A. J. G. (1986) JGRQl, 9277. [6] Stevenson D. J. (1986) GRL,13, 1149. [7] Jurewicz S . R. and Jones J. H. (1994) LPSC XXV,2,653 [8] Shannon M. C. and Agee C. B. (1995) EOS,76, F698. [9] Herpfer M. A. and Larimer J. W. (1993) Meteoritics,28,362. [lo] Herpfer M. A. (1992) Ph.D. ASU.Work performed under the auspices of D.O.E. contract W-7405-Eng-48 AT L LNJL