Experimental procedures were designed to test the most reliable and consistent methods to study the particle size distributions and characterize the texture of Leg 201 samples with the combined use of a Beckman-Coulter LS 12 320 laser particle sizer and visual verification of compositions and sizes of the particles with optical petrography. The experiments were designed to address the problems involving the pretreatment methods necessary to disaggregate the sediments and their possible effects on sediment grains either by destroying or creating new particles. These problems include the effects of grinding of the dried and consolidated sample during subsampling and sample pretreatment and the effects of sonication on the sample in the liquid module of the particle sizer during the analysis. Verification of grain types, visual calibration of grain sizes, and identification of the effects of the pretreatment methods were carried out with petrographic microscopy analysis of smear slides. In order to characterize the relative contribution to the grain size distributions of the two main biosiliceous and biocalcareous sedimentary components, analyses were performed on both the bulk and noncarbonate fraction of the samples (the latter was obtained by reacting the sediment with HCl; see below).
The biogenic calcareous and siliceous sediments of Leg 201 have a chalklike appearance because they have been drying for a few years since they were collected during Leg 201 in 2002. During subsampling, the core sediments break in irregular centimeter- to millimeter-size chunks that need to be disaggregated prior to particle size analysis. Subsampling was done by collecting a few grams of the indurated sediment and then grinding and powdering the subsample in a mortar with a ceramic pestle for a short time (~30 s; maximum = 1 min). The subsample was then separated in two aliquots (0.5 and 1.0 g) for analysis of bulk and noncarbonate fractions, respectively (see below). Given the very small size of the particles composing Leg 201 sediments (mean diameter is generally <60 µm) (Tables T1, T2), the bias introduced by grinding is probably negligible. However, the possible destructive effects of grinding on the pelagic particles were tested by analyzing the grain size distributions of the test samples over increasing amounts of grinding time (from 30 s to 5 min). The results of the tests supported our assumption that grinding of the samples for a short time (<1 min) does not alter the grain size distributions, which are significantly affected only by prolonged (>3 min) grinding.
Organic carbon is another component that can bind grains and forms particle aggregates that need to be dissolved with hydrogen peroxide before analysis. However, organic carbon in the eastern equatorial Pacific sediments is generally low (average = ~0.25 wt%) and rarely exceeds 1 wt% (D'Hondt, Jørgensen, Miller, et al., 2003), so treatment with H2O2 to oxidize organic matter was not necessary. Other factors however, may potentially affect the results of grain size analyses, including the destruction of clay minerals by HCl treatment (Stuut, 2002), corrosion of coccoliths by deionized water (Andruleit, 2000), and air bubbles in the liquid solution of the particle sizer.
After grinding, further disaggregation of the samples was first achieved by stirring the diluted sediment sample in a flask for a few seconds. Then, during particle size analysis, disaggregation was caused by the energy of the pump in the sample cell and, finally, by sonication. In order to establish the most reliable procedure for disaggregating the sediment while in the liquid module of the laser particle sizer and obtain reproducible particle size results that match visual observation on smear slide, several sonication tests were performed. Sonication is the most effective tool for the destruction of particle aggregates but can also add error to the measurements by creating new, smaller particles as a consequence of the fragmentation of whole fragile microfossil tests. For the biogenic sediments from the Walvis Ridge, Stuut et al. (2002) used 15 s of ultrasonic dispersion to prevent the fragmentation of delicate foraminifer tests. The optimal sonication time for carbonate-rich adhesive lake sediments was found to be 45 s (Murray, 2002).
The variations of mean diameter and mode with increasing sonication time for two test samples, diatom and nannofossil oozes (Samples 201-1226B-31X-3, 0 cm, and 24H-3, 0 cm, respectively) are shown in Figure F2A and F2C. The figure also shows the particle size distribution of the test samples at 0 s, 30 s, and 2 min sonication times (Fig. F2B, F2D). The results indicate that most of the decrease of the mean size occurs within the first few seconds of sonication, and after ~30 s the mean size decreases at a lower rate. However, the statistical results in Tables T1 and T2 show that on average there is ~50% reduction of mean size after 30 s of sonication in the bulk fraction and only a minor change in the noncarbonate fraction. Sonication is probably more effective in the bulk rather than in the noncarbonate fraction because of the strongly adhesive quality and highly charged molecules of carbonate-rich material (Murray, 2002). According to the results of the sonication experiments and the visual observations of the sediment samples under petrographic microscope, a sonication time of 30 s was chosen because it represents the best balance between the need to prevent the formation of aggregates in suspension and the need to prevent the formation of new particles due to the fragmentation of fragile biogenic debris.
Subsampling of the solutions for the laser particle analysis was done with a pipette (diameter = >2 mm) while energetically stirring the flask to resuspend the sediment and ensure random sampling. Increasing amounts of the sediment solution were then added to the aqueous module of the particle sizer until obscuration values of 8%–12% and PIDS obscuration values of 48%–52% were obtained. Obscuration is the percentage or fraction of light that is attenuated because of extinction (scattering and/or absorption) by the particles and is also known as optical concentration.
The grain size analysis of the noncarbonate fraction was achieved by reacting the total carbon with HCl. Preparation of the noncarbonate fraction was done in conjunction with measurement of the weight percent total inorganic carbon (TIC). For TIC analyses, ~1 g of dry sediment was mixed with 10 mL of HCl in a sealed chamber, the pressure developed by the reaction was measured for 60 s, and the peak value was reported in pounds per square inch. Pressure values were converted to weight percent TIC by comparison with a standard curve. For more details on the gasometric method (carbonate "bomb"), see Schink et al. (1979). After reading the pressure, the lid of the chamber was removed and the remaining sample was resuspended in distilled water to increase the pH, in order to avoid formation and precipitation of reaction products. The sample was then rapidly moved to the sample cell of the aqueous module and measured for grain size.
Instrument settings during operations were as follows:
Deionized water was used to supply the liquid module. The optical model chosen for grain size determination is the default Fraunhofer model, based on the Fraunhofer theory of light scattering.
For all samples analyzed, grain size was also measured prior to sonication. The results of the analyses are reported in the "Appendix," and the statistical summaries for Sites 1225 and 1226 are presented in Tables T1 and T2, respectively.
Data interpolation and statistical analyses were obtained with the laser particle sizer proprietary software (Beckman Coulter Inc., 2003). Because all samples analyzed tend to log-normal grain size distributions in the 0.04 µm to 2 mm spectrum, geometric rather than arithmetic statistics were applied to the values obtained by the logarithmically spaced size channels of the particle sizer. The statistical results of the grain size analyses of the samples from Sites 1225 and 1226 listed in the "Appendix" include the standard ODP sample header. Weight percent TIC determined with the gasometric method is also reported. Particle size data include the mean size and mode for both bulk and noncarbonate fractions for two sonication times (0 and 30 s). For the samples analyzed after 30 s of sonication, further statistical data listed include standard deviation (SD) and three percentiles (10%, 50%, and 90%).