INTRODUCTION

This paper deals with a new nondestructive technique for evaluating the igneous stratigraphy, structure, and composition of cores of gabbroic rock. I illustrate its utility by comparing the measurement to observations on cores recovered from Ocean Drilling Program (ODP) Hole 735B, Southwest Indian Ridge, during Leg 176 and to the bulk compositions of the rocks. The technique is one previously applied to problems in high-resolution stratigraphy of marine sediments obtained by hydraulic piston coring from JOIDES Resolution. It provides results of comparable detail and quality on continuously cored gabbro.

During ODP Leg 118, downhole logging of magnetic susceptibility in Hole 735B recorded variations between gabbros rich in the magmatic oxides, ilmenite, and magnetite, which produced sharp peaks, and oxide-poor olivine gabbros and troctolites, which produced the intervening lows (Pariso et al., 1991). The peaks, or high-susceptibility spikes, were most numerous near the top and bottom of the 490 m logged, and there was one longer interval between ~220 and 280 meters below the seafloor (mbsf) with systematically high magnetic susceptibility (Fig. F1). The pattern of peaks accords well with the placement of oxide gabbros in descriptions of the recovered core (Shipboard Scientific Party, 1989).

The logging tool sensed the presence of magnetite. Pure magnetite has a very high magnetic susceptibility ( = 3 SI units) (Collinson, 1983; Hunt et al., 1995), as does pyrrhotite, whereas the susceptibilities of other ferrimagnetic minerals likely to be in the rocks (e.g., ilmenite and hematite) are one to three orders of magnitude smaller, and silicate minerals have = 0.001 SI. Of the oxide minerals and sulfides, ilmenite ( = ~0.009 SI) is the only one besides magnetite that is volumetrically significant in any of the rocks. It is always intergrown with magnetite and in many rocks with tiny amounts of globular magmatic sulfide, which includes pyrrhotite (Natland et al., 1991; Natland and Dick, 2001). Ilmenite and magnetite together comprise between ~2% and 30% of the mode of oxide gabbros and gabbronorites, ~1%-2% of disseminated oxide gabbros and gabbronorites and <1% (oftentimes <<1%) of olivine gabbros, troctolitic gabbros, and troctolites. Only the latter rocks may contain more magmatic pyrrhotite than magnetite. Measurements of magnetic susceptibility on minicores of these lithologies decrease by about two orders of magnitude in that order.

There are at least three types of magnetite in the gabbros (Natland et al., 1991; Shipboard Scientific Party, 1999b). The most abundant is magmatic in origin. It is actually a solid solution of magnetite (Fe²⁺Fe₂³⁺O₄) and ulvöspinel (Fe₂²⁺TiO₄) or Mt-Usp_ss (Natland et al., 1991). The ilmenite intergrown with it is a solid solution of ilmenite (FeTiO₃) and hematite (Fe₂O₃), or Il-Hem_ss, and it is usually several times more abundant than magnetite in the intergrowths. The two minerals are present as platelike intergrowths; rarely is ilmenite exsolved from magnetite (Shipboard Scientific Party, 1989).

The second variety is almost pure magnetite. It is much finer grained and forms during alteration along with secondary amphibole and layer-lattice silicates replacing olivine and pyroxenes. A third variety is also nearly pure Mt, and it formed by exsolution from olivine, pyroxenes, and plagioclase during high-temperature subsolidus recrystallization of the rocks (Shipboard Scientific Party, 1999b). Such magnetite is not volumetrically important in oxide gabbros, but it may be the most important magnetic mineral in olivine gabbros and troctolites, especially in the usual case where the rocks are very little altered. Because magmatic oxides crystallize late during the differentiation of basaltic magma and often have particular relationships to gabbroic textures (Natland et al., 1991; Natland and Dick, 1996, 2001), any systematic nondestructive measure of the abundance of magmatic oxides is potentially a very useful petrogenetic index.

Although there was no magnetic logging during Leg 176, magnetic susceptibility was routinely and systematically measured on recovered rock, which amounted to >86% of the 1003.2 m cored during the leg, using a Bartington MS2C sensor integrated with the multisensor track (MST). This is the same instrument that is used to detect fluctuations in magnetic susceptibility of marine sediments obtained by piston coring. During Leg 176, measurements were almost always at 4-cm spacing and were made as the rock was carried along a moving track through the instrument, in the same way and at the same scale as measurements on sediments. The measurements were obtained on whole-round core immediately after the pieces of rocks were spaced in half liners, thus are directly related to curated depth, the measure by which cores are described, photographed, and sampled. Additional measurements were made on 339 minicore samples at an average spacing of ~3 m. The MST, on the other hand, was able to resolve local concentrations of magmatic iron-titanium oxides on the scale of only a few centimeters, and these could later almost always be identified visually in the cores (Shipboard Scientific Party, 1999b). The measurements show that the pattern of spikes originally seen in the downhole log of Leg 118 persists into deeper rock and exists on the scale of the individual piece of rock measured in a single section of core. Several hundred large and small spikes in magnetic susceptibility, each spike only a few centimeters to a few tens of centimeters in length, were thus recorded between 504 and 1008 mbsf. MST magnetic susceptibility data therefore provide a novel and potentially extremely precise means to evaluate gabbro lithology in cores. The recovery during Leg 176 was also sufficiently high that the measurements provide a statistically valid assessment of the proportion of rocks with high magnetic susceptibility, thus with high concentrations of iron-titanium oxides, throughout the long section cored.

In this paper, I illustrate some of the ways in which spikes in magnetic susceptibility coincide with lithologic features of the core. I also demonstrate that there are strong correlations between magnetic susceptibility and chemical attributes of the rocks related to the abundance of ilmenite and magnetite, thus that magnetic susceptibility is a useful and quantitative geochemical log. It can even provide information about the dispersed and low, but nevertheless varying, concentrations of magnetite in olivine gabbros and troctolites. Combining the logging results from Leg 118 with the MST data from Leg 176 provides the only nondestructive log pertaining to the bulk composition of the rocks obtained over the entire core. It provides a means of refining estimates of the bulk composition of the entire section.