to 65°2
at 0.5°2
per minute. Sample holders were rotated during analyses to minimize preferred
orientation.Thirty-one bulk sediment samples between 25.38 and 29.49 meters below seafloor (mbsf) were collected from Site 1063. The sampling interval is typically 0.1 m but in some cases is up to 0.5 m. These bulk samples were oven dried at 105°C for at least 24-36 hr to remove pore water. Approximately 2 g of each sample was crushed and homogenized for 10 s in a Rocklabs mini-mill with a tungsten carbide mill head and holder. A small amount of sample (~0.1 g) was smeared onto a glass disc using distilled water and subsequently dried. Care was taken to ensure the glass disc was completely covered with sample.
All crushed and
homogenized samples were analyzed for mineralogy using a Siemens D-5000 X-ray
diffractometer at the James Cook University Advanced Analytical Centre (JCU-AAC).
The diffractometer was fitted with a copper tube, slit system, and postsample
graphite monochromator. All samples were run at 40 kV and 30 mA and from 1.3°2
to 65°2
at 0.5°2
per minute. Sample holders were rotated during analyses to minimize preferred
orientation.
Quantification of the resulting X-ray diffraction (XRD) data was by means of a "Rietveld" synthesis using the software package Siroquant, which quantifies the phases present by fitting a synthetic XRD profile to the observed pattern (Taylor, 1991). The synthetic profile is calculated by constructing diffraction patterns for each phase thought to be present in the sample from the respective fundamental crystal structure data.
Amorphous (biogenic) silica was quantified using a modified cristobalite diffraction model, whereby the cristobalite peak width parameter (w) was empirically modified to fit the degree of crystallinity evident in the diffraction pattern. The accuracy of this method to estimate silica was assessed by quantifying known mixtures of chromatography grade silica gel and quartz. Four quartz and silica gel mixtures were crushed and homogenized for 10 s in a tungsten carbide ring mill and smeared on glass discs. Care was taken to ensure the glass disc was completely covered. Results of these analyses of quartz-silica gel mixtures are given in Table T1. As silica gel may have different spectral properties from biogenic silica, it is unclear whether this accuracy applies to estimations of the abundance of biogenic silica in samples from Site 1063. However, the amount of biogenic silica in any given sample can be constrained from concentrations of major element oxides (see below).
Small (1-2 g) splits of the bulk samples dried and crushed for XRD analyses were analyzed for major element oxide concentrations by X-ray fluorescence (XRF). Powdered samples were roasted at 1000°C for about 14 hr. Approximately 0.6 g of each roasted powder was then mixed with Norrish Hutton XRF flux and fused into a glass disc. Major element oxide concentrations were obtained using a Siemens SRS-3000 XRF spectrometer at JCU-AAC by comparison to concentration curves defined by analyses of 15 geological standards. Repeated analyses of multiple pressed glass discs of a common geological standard (AGV-1) each have concentrations of all major element oxides to within 2% of certified values (Gladney et al., 1992).
Particle size was determined on selected bulk samples between 25.8 and 27.2 mbsf. The chosen depth interval corresponds to a zone of low bulk density and high P-wave velocity. All of these samples were examined with a Malvern Mastersizer X laser particle size analyzer in the James Cook University Marine Geophysical Laboratory (e.g., Woolfe et al., 1998). Use of this instrument involves adding suspended sediment to a sample chamber, which is then pumped via a recirculating system through a narrow glass cell of 18 mm diameter and 2.4 mm width. As particles pass through the cell they induce forward scattering of a He-Ne laser beam. Particle size is quantified from the diffraction of the laser beam.
Two procedures for preparing samples were used for determining particle size. Samples were initially disaggregated using an ultrasonic probe on the Mastersizer particle size analyzer. However, repeated measurements of a single sample over time indicated that a stable size distribution could not be reached for Bermuda Rise sediment using this method (Fig. F3). In particular, the amount of sediment between 10 and 200 µm decreases over time. Thus, an alternative procedure was followed. Approximately 0.6 g of wet sediment was suspended in 40 mL of distilled water in a screw cap container. The containers were placed inside a rotating cylinder for 12 hr to gently agitate the sediment. Repeated analyses of the same sample shows that results are reproducible with the same sample preparation (Fig. F4).
