DISCUSSION

According to Craig and Kullerud (1969), the sulfide assemblage of pyrrhotite, chalcopyrite, troilite, and pentlandite is typical of igneous sulfides segregated from mafic magmas that subsequently underwent lower temperature reequilibration. The presence of troilite, which is not commonly reported in mafic igneous rocks but which is present throughout the entire section sampled during both Legs 118 and 176 (with the exception of the upper 150 m or so), indicates that the pyrrhotite was originally quite Fe rich and underwent exsolution at low temperature (Kissin and Scott, 1982) to troilite plus pyrrhotite. Toulmin and Barton (1964) argue that troilite exsolution from Fe-rich pyrrhotite indicates lower sulfur fugacity than in assemblages that contain only relatively Fe-poorer pyrrhotite; however, in these samples troilite is present, associated with both Fe-rich and Fe-poor pyrrhotite (see Figs. F4, F5).

One method used in this investigation to target sulfide-bearing intervals that might contain PGE was to identify intervals that show anomalously high values of pathfinder elements that have partitioning behavior similar to PGE. Figure F10 illustrates the concentration with depth in the core of two of the most useful of these pathfinder elements, Cu and Ni. Cu and Ni are both chalcophile and siderophile, partitioning strongly into sulfide phases in a magmatic system. Most of these data are from the shipboard data sets of Legs 118 and 176 (Robinson, Von Herzen, et al., 1989; Dick, Natland, Miller, et al., 1999). New bulk chemical analyses resulting from this study are presented in Table T2. Ni and Cu average 110 and 65 ppm, respectively, throughout the section, and both show an overall slight increase in concentration with depth.

Ni concentration relative to depth shows many more high-concentration spikes than does Cu. There are two possible explanations for this phenomenon. Figure F11 is a plot of bulk rock Ni and Cu for all analyses available from Hole 735B. These have been subdivided into oxide-bearing gabbros, gabbros, and olivine gabbros, based on petrography. All oxide-bearing gabbros, regardless of grain size or olivine content, have been grouped together for this illustration. Virtually all of the oxide-bearing gabbros show depleted Ni contents at any given Cu content. Presumably this is because these gabbros precipitated from differentiated liquids where olivine had sequestered most of the available Ni. The gabbros and olivine-bearing gabbros show a pattern with coincident increase in Cu and Ni but substantial variability. In some of the samples, Ni values above the generally increasing Cu-Ni line can be attributed to olivine accumulation. Average Ni abundance in olivine reported from Leg 118 postcruise research (Ozawa et al., 1991) is 100 to 300 ppm in gabbros and oxide-bearing gabbros but exceeds 1500 ppm in troctolites. Where the modal olivine abundance in troctolites approaches 50%, high bulk rock Ni contents are possible. In all the samples examined from the Leg 176 suite, however, it appears that in each case the high bulk Ni compositions are due to a marked change in the ratio of pentlandite to pyrrhotite in the sulfide assemblage. In general, pentlandite comprises <10% of the total sulfide minerals (see Figs. F4, F5). In each interval where there is a high Ni concentration, the proportion of pentlandite increases, with pentlandite accounting for as much as 90% of the sulfide mineral assemblage (Fig. F12). Pentlandite in these samples contains >30 wt% Ni, and this can account for enrichments as high as 800 ppm in bulk Ni composition.

Cu and Ni bulk concentrations also mimic lithostratigraphic variation based on magnesium number (compare Figs. F3 and F10). The variability of Cu and Ni concentration is fairly constant below 500 mbsf, except in the interval from 960 to 1170 mbsf, where compositional range is very limited. As noted earlier, the lower contact of this interval is coincident with a major shift in pyrrhotite composition, a sharp decrease in magmatic foliation, and a sharp decrease in magnetic susceptibility (Dick, Natland, Miller, et al., 1999). All of these indicate that the gabbroic rock below 1170 mbsf precipitated from a different magma batch than the gabbro that is higher upsection.

In those sulfides analyzed by electron microprobe, pentlandite is the host for nearly all of the detectable Ni and Co. According to Mazdab and Force (1998), comparison of Co/Ni ratios in sulfides readily differentiate Fe oxide deposits (Co/Ni > 1) from magmatic Fe-Ti oxide or magmatic immiscible sulfide Ni-Fe-Cu systems (Co/Ni < 1). All samples in this study have a Co/Ni > 1, indicating that they are magmatic.

Although the initial focus of this investigation was PGE mineralization, only a few of the samples analyzed proved to have concentrations above a background of ~0.4 ppb total PGE and Au (Table T3). Concentrations significantly higher than this background value are present in the interval between 500 and 520 mbsf, and 744, 1192, and 1430 mbsf. The only petrographic or geochemical similarity among these samples is the presence of pentlandite with pyrrhotite and chalcopyrite. These samples do not have universally distinct proportions of these phases, and several other samples have this sulfide assemblage and do not have elevated PGE. However, each of these samples does have distinct characteristics from the routine gabbro sampled in Hole 735B. The interval between 500 and 520 mbsf (Samples 176-735B-90R-1, 123-125 cm; 90R-4, 42-45 cm; 90R-6, 72-77 cm; and 176-735B-91R-1, 106-111 cm) all have much higher than average bulk Ni composition and contain more abundant pentlandite than most of the samples examined. The sample from 744 mbsf (Sample 176-735B-123R-4, 106-111 cm) has the highest analyzed value of Pd and contains no detectable Pt. This sample contains armored polyphase sulfide globules with the most convincing igneous textures observed (see Fig. F4). Another petrographic distinction of this sample is that throughout the section there is, in general, an antithetic relationship between the abundances of olivine and sulfide phases. As the amount of olivine increases, the abundance of primary sulfide minerals declines. This sample, however, contains both abundant olivine and a relatively high abundance of sulfide minerals. Shipboard core descriptions (Dick, Natland, Miller, et al., 1999) identified this interval as a microgabbro, rich in olivine and sulfides, with an intrusive contact with the surrounding pegmatitic-textured gabbro. Although structurally a late intrusive event, based on the abundance of olivine and primary sulfides, this interval must represent a relatively primitive magma. The sample from 1192 mbsf (Sample 176-735B-175R-1, 115-120 cm) is coincident with the geochemical boundary marked by the radical change in primary sulfide composition. The sample from 1430 mbsf (Sample 176-735B-203R-1, 11-13 cm) is from a 6-cm-thick chalcopyrite-, pentlandite-, and pyrrhotite-bearing clinopyroxenite and contains the highest abundance of PGE detected. A thin section from this sample was examined in an exhaustive search under reflected light at high magnification with a petrographic microscope and with backscatter electron imaging on the electron microprobe, and no platinum-group element-bearing minerals were detected. However, all of the PGEs in this sample, if concentrated in a single grain, would require that the grain be only a few micrometers across.

NEXT