All 174 samples were freeze-dried and gently ground with a mortar and pestle. Three parameters of bulk sediment were then determined on some or all of the samples: (1) carbonate content, (2) biogenic silica content, and (3) scandium (Sc) concentration. Scandium concentrations were examined because Sc is a good proxy for the amount of aluminosilicate (clay) material in marine sediment (Schmitz et al., 1991; Dickens and Owen, 1996). Two selected samples were then examined for microstructure.
Calcium carbonate was determined using the standard carbonate bomb technique of Müller and Gastner (1971). Of the 129 samples analyzed, 35 were from within the upper hydrate zone and 94 were from above and below the zone. Each 1 g split of powdered sample was combined with 15 mL 3N HCl in a reaction vessel attached to a water filled burette. The CO2 released was measured and compared to a standard curve derived from linear regression of CaCO3 standards. A new curve was generated each day to account for variations in atmospheric pressure and room temperature. The precision of this method was within 1% to 2% as determined from known standards and replicate analyses.
Biogenic silica (opal) was determined on an atomic absorption spectrometer (AAS) with a heated graphite atomizer using a method modified from Strickland and Parsons (1968) and Eggimann et al. (1980). From a total of 110 samples, 47 were from within the upper hydrate zone and 73 were from outside the zone. About 50 mg of a powdered sample was heated in a solution of 1% Na2CO3 to leach the opal. The liquid was then diluted to within working range and analyzed for Si on the AAS. Silicon leached from clays was found to be inconsequential because the contribution from Al absorbance was negligible. The Si concentration was determined by comparing the absorption caused by Si to a linearly regressed standard curve derived from 25 mL standard solutions neutralized with 1 mL concentrated HCl. Percent opal was determined stoichiometrically. Analytical precision, based on replicate analyses, was usually ± 5% for Si.
Scandium was determined on
101 samples by instrumental neutron activation analysis (INAA) using the nuclear
reactor and counting facilities at the Phoenix Memorial Laboratory, University
of Michigan. Sample preparations and analytical procedures were the same as
those used by Dickens and Owen (1995), with the exception of a new
data-processing system that allows higher resolution measurements. Analytical
precision (1
)
is within 3%, and total estimated errors (1)
including a range of random errors is within 5%. The precision and accuracy were
evaluated by analyzing nine samples of NBS-SRM-679 (brick clay) in batches with
Blake Ridge samples. Mean Sc concentration for the analyses of NBS-SRM-679 was
22.7 ppm with a standard deviation of 0.39 ppm. This compares to reported
average concentrations for Sc in NBS-SRM-679 of 22.5 ppm (Gladney
et al., 1987) and 23.1 ppm (Dickens and Owen, 1995).
Samples 164-994C-24X-1, 39-41 cm, and 164-994C-23X-4, 40-42 cm, were impregnated with a resin/ether plastic mixture using the method developed by J. Mauk at the University of Michigan (J. Mauk, pers. comm., 1998). Samples were immersed in the plastic mixture and placed in an impregnation chamber. The chamber was then evacuated to cause the plastic to impregnate the permeable samples, and the samples were allowed to harden. The representation of in situ porosity is preserved because the sample can then be cut into thin sections without causing portions of the sample to decompose into individual grains and cause erroneous pore spaces. Thin sections were then carbon coated. A Hitachi SEM equipped with a backscatter-electron (BSE) imaging system was used to study in situ porosity and microporosity of these samples.
Porosity calculations were made from ODP bulk porosity values (Paull, Matsumoto, Wallace, et al., 1996). We fit a smooth curve through the ODP porosity data to the bottom of the hole to predict the porosity at each of the sample depths because samples are offset in depth from porosity measurements. Expected porosity is based on a log curve through all porosity data at Site 994 to account for the exponential decay of porosity because of sediment compaction. Normalized porosity (observed porosity - expected porosity) was calculated to subtract the compaction effect because there are different sediments further downcore. However, we found that normalized porosity made a negligible difference on our results.
