Previous   |    Next

doi:10.2204/iodp.proc.322.202.2013

Methods

Calculations of mineral abundance

Marine sediment samples can be analyzed by XRD using a variety of techniques. For example, the presence of a specific detrital and/or authigenic mineral can be detected easily through visual recognition of characteristic (hkl) peak positions. It is more challenging, however, to estimate the relative abundance of a mineral in bulk sediment or in the clay-size fraction with meaningful accuracy (e.g., Moore, 1968; Heath and Pisias, 1979; Johnson et al., 1985). The most common approach for analyzing clays in marine geology has been to multiply the Biscaye (1965) weighting factors by the peak areas of basal reflections and normalize to 100% (McManus, 1991). Errors can be substantial, however, and accuracy of calculated values is affected by the absolute abundance by weight of each mineral in the mixture (Underwood et al., 2003). XRD results also change depending on sample disaggregation technique, chemical pretreatments, particle size separation, crystallinity and chemical composition of minerals, peak-fitting algorithms, and the degree of preferred orientation of crystallites (e.g., Moore and Reynolds, 1989; Ottner et al., 2000). Even though data reproducibility might be very good, accuracy is usually no better than ±10% unless the analytical methods include calibration with internal standards, use of single-line reference intensity ratios, and some fairly elaborate sample preparation steps to create random particle orientations (Środoń et al., 2001; Omotoso et al., 2006).

One goal of NanTroSEIZE is to obtain internally consistent, semiquantitative estimates of mineral abundance in the clay-size fraction for a large number of samples. To accomplish this, we use a matrix of singular value decomposition (SVD) normalization factors, as documented in full detail by Underwood et al. (2003). Figure F3 shows representative X-ray diffractograms for two clay-size aggregates from the Shikoku Basin. The matrix of SVD factors (Table T1) is applied to the integrated areas of a broad smectite (001) peak centered at ~5.3°2θ (d-value = 16.5 Å), the illite (001) peak at ~8.9°2θ (d-value = 9.9 Å), the composite chlorite (002) + kaolinite (001) peak at 12.5°2θ (d-value = 7.06 Å), and the quartz (100) peak at 20.85°2θ (d-value = 4.26 Å). Average errors for the standard mineral mixtures used to calibrate this method are approximately 3% for smectite, 1% for illite, 2% for chlorite, and 1.4% for quartz (Underwood et al., 2003). Because of the nearly total overlap between the kaolinite (001) and chlorite (002) reflections, we first calculate that relative abundance as undifferentiated chlorite + kaolinite, and then solve for the proportion of each mineral using the double peak at ~25°2θ (Fig. F3) and a refined version of the Biscaye (1964) method, as documented fully by Guo and Underwood (2011). Analysis of standard mineral mixtures shows that the average error for the chlorite/kaolinite ratio is 2.6%. To provide an estimate of the abundance of individual clay minerals in the bulk mudstone, we also multiply each relative percentage among the clay minerals (i.e., excluding quartz) by the weight percent of total clay minerals from shipboard bulk powder XRD analyses of co-located “cluster” specimens (e.g., see the “Site C0011” chapter [Expedition 322 Scientists, 2010a]). To facilitate comparisons with many of the other published data sets from the region, data tables include weighted peak area percentages for smectite, illite, chlorite, and kaolinite using Biscaye (1965) weighting factors. These values are relative percentages and should be regarded as semiquantitative.

To characterize the extent of clay diagenesis, we used the saddle/peak method of Rettke (1981) to calculate the percent expandability of smectite and illite/smectite (I/S) mixed-layer clay. This method is sensitive to the proportions of discrete illite (I) versus I/S mixed-layer clay. Our calculations follow a curve for 1:1 mixtures of I and I/S. A complementary way to calculate the proportion of illite in the I/S mixed-layer phase is based on the position (d-value) of the (002/003) peak (following Moore and Reynolds, 1989) after correcting the diffractogram peaks for misalignment of the detector and sample holder. We also report values of illite crystallinity index as the peak width measured at half height (Δ°2θ) for the (001) reflection.

Sample preparation

Isolation of clay-size fractions starts with air drying and gentle hand-crushing of the mud/mudstone with a mortar and pestle, after which specimens are immersed in 3% H₂O₂ for at least 24 h to digest organic matter. We then add ~250 mL of Na hexametaphosphate solution (concentration of 4 g/1000 mL distilled H₂O) and insert the beakers into an ultrasonic bath for several minutes to promote disaggregation and deflocculation. These steps are repeated until disaggregation is complete. Washing consists of two passes through a centrifuge (8200 revolutions per minute [rpm] for 25 min; ~6000 g) with resuspension in distilled-deionized water after each pass. After transferring the suspended sediment to a 60 mL plastic bottle, each sample is resuspended by vigorous shaking and a 2 min application of a sonic cell probe. The clay-size splits (<2 µm spherical equivalent settling diameter) are then separated by centrifugation (1000 rpm for 2.4 min; ~320 g). Oriented clay aggregates are prepared using the filter-peel method (Moore and Reynolds, 1989) and 0.45 µm membranes. The clay aggregates are saturated with ethylene glycol vapor for at least 24 h prior to XRD analysis, using a closed vapor chamber heated to 60°C.

X-ray diffraction parameters

The XRD laboratory at the University of Missouri (USA) utilizes a Scintag Pad V X-ray diffractometer with CuKα radiation (1.54 Å) and a Ni filter. Scans of oriented clay aggregates are run at 40 kV and 30 mA over a scanning range of 3°–26.5°2θ, a rate of 1°2θ/min, and a step size of 0.01°2θ. Slits are 0.5 mm (divergence) and 0.2 mm (receiving). The digital data are processed using MacDiff software (version 4.2.5) to establish a baseline of intensity, smooth counts, correct peak positions offset by misalignment of the detector (using the quartz [100] peak at 20.95°2θ; d-value = 4.24 Å), determine peak intensity (counts/step), and calculate integrated peak areas (total counts). This program also calculates peak width at half height (Δ°2θ).

Top of page   |    Previous   |    Next