Trace-element and isotopic compositions of some troctolites and gabbros collected from the Mid-Cayman Rise (MCR) spreading center are consistent with formation of these rocks in open systems. Contents of incompatible trace elements (Zr, LREE) are too high for their incorporation only in the olivine, plagioclase, and Ca-rich clinopyroxene that dominate these rocks, assuming crystallization from liquids similar to those that formed the glasses also collected from this region. Additionally, mineral separates exhibit heterogeneity in their Nd and Pb isotopic compositions within a few cubic centimeters of one rock. The presence of exotic material in these rocks is particularly surprising as the major minerals in most gabbroic rocks from the MCR are essentially unzoned in major element contents.
We are conducting detailed analyses of some of these rocks to try to locate the exotic material. The construction of one mafic troctolite will be discussed in some detail. This rock is dominated by olivine (Fo85.2±0.6) and plagioclase (An72.3±0.5), with a few large grains of Ca-rich clinopyroxene (Mg#=87.2±0.8). Interstitial grains include ilmenite, magnetite, apatite, and rare baddeleyite. These grains are usually located in regions occupied by minerals of hydro-thermal origin, the most abundant of which are clinozoisite, amphibole, serpentine, chlorite, and various iron oxides. The small apatite grains (none exceed 20 mm) are almost all enclosed in 30-60 mm grains of clinozoisite that appear to be continuous with stringers along grain boundaries and, presumably are indicative of the movement of hydrothermal fluids through the rocks after their solidification.
The liquid that crystallized the primocrystic gabbroic minerals was relatively primitive. If similar to the liquids that formed the glasses in the region, it contained 200-300 ppm Zr and <0.3% P2O5. Model calculations show that the liquids that formed the MCR glasses were not saturated in iron-titanium oxide until they had evolved to much lower values of Mg# and An/(An+Ab). The apatite, iron oxides, and zirconia represent material incorporated from a different source. Three models for the origin of these materials are considered. They could have been introduced hydrothermally during the event that formed the hydrous minerals listed above. They could have crystallized from a silicate liquid related to the melt that formed the primocrysts but that had undergone enrichment of incompatible trace-elements by crystallization. Alternatively, these minerals formed from a mafic silicate liquid unrelated to the melt that formed the primocrysts.
The spatial association of the apatite and oxides with hydrous alteration products is suggestive that the former formed from or were mobilized by hydrothermal fluids. They may also be associated if hydrothermal fluids travel along pathways that were previously used by late-stage
silicate liquids. The apatite grains have F>Cl~OH, whereas enclosing clinozoisite grains (and all other analyzed grains of definitively hydrothermal minerals) have undetectably low halogen contents. Although apatite may concentrate halogens with respect to a coexisting fugitive phase, the stark contrast in site occupancies suggest that the apatite grains did not form in equilibrium with the clinozoisite grains but predate them.
Traverses from cores of plagioclase grains to the borders adjacent to the apatite-enclosing clinozoisite grains demonstrates a complete lack of chemical zonation in the mineral at the level of reproducibility of the electron probe. Backscattered electron maps of the grains also indicate the lack of heterogeneity in major elements. The hydrothermal fluids that penetrated these rocks did not alter primary phases except along grain boundaries. This is not suggestive of massive introduction of incompatible trace elements into these rocks by fluids. Modeling introduction of the exotic component by late-stage dioritic or more evolved liquids faces similar problems. If apatite and oxides formed in veins produced by such liquids, the veins would also have contained albitic plagioclase and more iron-rich olivine, pyroxene, or both.
Although plagioclase grains exhibit no major-element zonation, their contents of some minor and trace elements vary greatly. Ti contents show very large ranges. Plagioclase near the apatite grains have Ti contents that exceed those of the cores by a factor of 5 or more, whereas reproducibility is better than 10%. Ion-probe data shows that Ti and LREE contents are correlated in plagioclase in MCR gabbros and suggests that the variations within plagioclase are due to formation from two different liquids. We believe that the apatite and oxide grains and the bordering adcumulate grains formed from a silicate liquid with much higher contents of some incompatible trace elements than the liquids that formed the primocrysts. This low-abundance liquid percolated through the almost consolidated cumulate pile, reacting with primocrysts and crystallizing both adcumulate and intercumulate mineral grains.