: (1) air-dried, (2) glycolated (12 hr under vacuum in ethylene glycol at room temperature), and (3) heated (2 hr at 490°C) from 2.5° to 32.5°2.
During Ocean Drilling Program (ODP) Leg 184, a series of six sites was drilled in the South China Sea. Among them, Site 1146, located within a small rift basin on the midcontinental slope of the northern SCS (Fig. F1), was cored with a maximum penetration of 633 meters below seafloor (mbsf). This core contains hemipelagic sediments dating from the early Miocene to the Pleistocene. For this study, our analysis concentrated on the latest 2 m.y., corresponding to the upper 190 meters composite depth (mcd). The sediment is relatively homogeneous and consists of greenish gray nannofossil clay with an average of ~21 wt% carbonate content, and the age model was established using paleontological and magnetic data (Shipboard Scientific Party, 2000).
Clay mineralogy determination was performed by standard X-ray diffraction (XRD) on a set of 508 samples on the composite splice from Site 1146 from Holes 1146A, 1146B, and 1146C. According to the preliminary sedimentation rate calculated onboard and in order to keep an average sampling interval of 4 k.y., the sampling interval was averaged to 40 cm, reducing with decreasing sedimentation rate downcore. An extra set of 181 samples was later collected from the upper 60 m for refining the latest intervals. All samples were disaggregated in water and treated with 0.2-N hydrochloric acid to remove carbonates. During this operation, the suspension was kept agitated and the pH checked to avoid irregular or prolonged exposure to the acid. The decarbonated suspensions were washed with demineralized water to remove excess ions and to help disperse the clays. Stoke's Law was applied for the determination of settling time and depth to isolate the <2-µm size fraction. The clay fraction was concentrated in a centrifuge, and the resulting paste was spread into a calibrated recess cut into rounded glass slides.
Preparation and measurement were performed according to the routine procedures of Holtzapffel (1985). Briefly, a Philips PW 1710 diffractometer equipped with a Cu tube and Ni monochromator was used with the following settings: 40 kV, 25-mA generator tension/current, 1° fixed divergence slit, 0.1° receiving slit, with sample spinner on. All samples were scanned three times in step-scan mode with 1 s per step of 0.02°2: (1) air-dried, (2) glycolated (12 hr under vacuum in ethylene glycol at room temperature), and (3) heated (2 hr at 490°C) from 2.5° to 32.5°2.
The diffractometer is driven by a computer that gives peak intensities for each 0.02°2 interval. The spectra were then studied using the MacDiff software (Petschick, 1997). We used the peak surface calculated by the software as an indicator of clay-mineral amount. We realize that this technique does not permit the production of fully quantitative results, but rather results in semiquantitative mineral percentages. Repeated measurements indicate that replicates of the same sample give very similar results and that peak ratio calculations are very stable, thus lowering the uncertainties resulting from preparation and analysis. Measurements have been carried out on the basal reflections from the ethylene glycol XRD diagram, adjusted on the 4.26-Å quartz peak.
