Previous   |    Next

doi:10.2204/iodp.proc.322.205.2013

Laboratory testing methodology

Sample handling and preparation

Bulk material from Cores 322-C0011B-30R through 58R, corresponding to 580.4–865.9 mbsf and lithologic Units III and IV, was collected in 10 bags with 24.9 kg in total and shipped to the University of Texas (UT) at Austin (Texas, USA) laboratory. Table T1 lists the mass, core, depth, and lithologic unit of each bag of collected material. About one half of the total mass is from lithologic Unit III and the other half is from lithologic Unit IV. We then air-dried the bulk material and ground it in a ball grinder to destroy aggregates in order to start resedimentation with an unstructured, dispersed fabric. The entire ground material was sieved through a #60 mesh (0.251 mm) and then homogenized to a single batch. We added silt-size silica (MIN U SIL 40) from US Silica to the baseline batch of Nankai silty claystone in the following dry mass proportions of silty claystone to silica: 100:00, 88:12, 76:24, 64:36, 52:48, and 40:60. The silt we used is a crystalline, fine, ground, poorly sorted (well graded) silica. The range of particle sizes for the silica extends from 63 to 0.7 µm. The coefficients of uniformity and curvature, often used in geotechnical engineering, are 7 and 2, respectively.

Sample mineralogy

The mineralogic composition of the Nankai silty claystone was measured by Macaulay Scientific Consulting LTD in Aberdeen, UK. Both whole rock and <2 µm clay fraction analyses were performed by X-ray powder diffraction (XRPD). The bulk samples were dried at 105°C, wet ground (in ethanol) in a McCrone mill, and spray dried to produce random powders. The XRPD patterns were recorded from 2° to 75°2θ using cobalt Kα radiation. Quantitative analysis was done by a normalized full pattern reference intensity ratio method (Hillier, 2000). The XRPD patterns were referenced to patterns from the International Centre for Diffraction Database (ICDD). Expanded uncertainty using a coverage of 2 (i.e., 95% confidence, given by ±X^0.35, where X = concentration in weight percent [e.g., 30 wt% ± 3.3]). There may also be uncertainty for phases at the trace level (<1%) as to whether or not the phase is truly present in the sample. The bulk sample contains quartz (24 wt%), feldspar (16 wt%), and clay minerals (59 wt%) with lesser amounts of calcite, pyrite, and halite (Table T2; Fig. F2A).

The clay fraction <2 µm was obtained by timed sedimentation. It was prepared as oriented mounts using the filter peel transfer technique and scanned from 2° to 45°2θ in the air-dried state, after glycolation, and after heating to 300°C for 1 h. Identified clay minerals were quantified using a mineral intensity factor approach based on calculated XRPD patterns. Uncertainty for clay minerals in amounts >10 wt% is estimated as better than ±5 wt% at the 95% confidence level. The clay-size fraction is dominated by smectite (85 wt%) with lesser amounts of illite (11 wt%), chlorite (3 wt%), and kaolinite (1 wt%) (Table T2; Fig. F2B). The expandability of the illite/smectite (I/S) mixed-layer clay was estimated from the data in table 8.3 in Moore and Reynolds (1997), which have been tabulated from calculated diffractograms for each specific composition, and is ~80%.

The bulk mineralogic composition of the homogenized, resedimented Nankai silty claystone is in agreement with the shipboard X-ray diffraction (XRD) measurements of bulk powders (Underwood et al., 2009). About 24 wt% quartz and 16 wt% feldspar were measured in the Nankai silty claystone, whereas the shipboard analyses measured on average 18 wt% and 11 wt%, respectively (Underwood et al., 2009). These slightly increased fractions of quartz and feldspar in the Nankai silty claystone are consistent with a 10 wt% decrease in total clay minerals compared to the shipboard samples (Underwood et al., 2009; Underwood and Guo, in press). The abundance of smectite in the bulk Nankai silty claystone is 45 wt%, which is on average equal to the smectite abundance in the bulk mudstone at Site C0011 (Underwood and Guo, in press).

The abundances of clay minerals in the clay-size fraction of the homogenized, resedimented Nankai silty claystone are consistent with XRD measurements on the clay-size fraction of the mudstone at Site C0011 as presented by Underwood and Guo (in press). Smectite is the dominant clay-size mineral. However, its abundance is higher in the Nankai silty claystone (85 wt%) than in the mudstone, where smectite averages 59 wt% in Unit III and 68 wt% in Unit IV (Underwood and Guo, in press). Similarly, values of illite (11 wt%) and chlorite + kaolinite (4 wt%) for the Nankai silty claystone are lower than for the mudstone, where illite averages 27 and 20 wt%, and chlorite + kaolinite average 4 and 7 wt% for Units III and IV, respectively (Underwood and Guo, in press). However, it has to be noted that these relative mineral abundances in the clay-size fraction of the mudstone also include clay-size quartz as opposed to only clay minerals in the Nankai silty claystone. The 80% I/S expandability of the Nankai silty claystone is consistent with an average I/S expandability of 77% of the mudstone (Underwood and Guo, in press).

Atterberg limits

We measured Atterberg limits such as the liquid limit (LL), plastic limit (PL), and plasticity index (PI) on the mudstone mixtures in accordance to American Society for Testing and Materials (ASTM) Standard D4318-05 (ASTM International, 2005). Atterberg limits are used to characterize the fine-grained fraction of soils and, together with other soil properties, to correlate with engineering behavior such as compressibility, hydraulic conductivity, intrinsic permeability, and shear strength, for example. Liquid and plastic limits are water contents that separate different consistency states from each other. We performed the multipoint LL method (ASTM International, 2005) using a hand-operated LL device, also called a Casagrande cup. We determined the PL by the hand method (ASTM International, 2005). PI is the range of water content over which the soil behaves plastically, or numerically, the difference between LL and PL.

Particle size analysis

After consolidation to 21 MPa, we ground all six samples with a mortar and pestle and determined the particle size distributions using the hydrometer method in general accordance with ASTM D422-63 guidelines (ASTM International, 2007) and Sawyer et al. (2008). Two hydrometer tests performed on the bulk Nankai silty claystone before compression produced almost identical particle size distribution curves than the hydrometer test performed after compression, indicating no mechanical effects on the particle size distribution. Therefore, we only performed hydrometer tests on the exact same specimens that were resedimented and uniaxially consolidated. The hydrometer method uses a suspension of sediment and water, which is thoroughly mixed, after which particles settle out of the water column according to Stoke’s law. We added 5 g of dispersing agent (sodium hexametaphosphate) to each suspension of 1000 mL. A hydrometer measures the specific gravity (G_s) of the suspension at known depths below the air-suspension interface and at known times. From the hydrometer tests, we obtain the particle diameter at a specific time and depth and the percentage of the original sample mass still left in suspension.

Resedimentation

We prepared six sediment samples in the laboratory using the resedimentation method (Santagata and Kang, 2007; Sheahan, 1991; Schneider, 2011) that was developed at the Massachusetts Institute of Technology and simulates natural sedimentation under controlled stress conditions. The dried, ground, and homogenized Nankai silty claystone was mixed at a water content of 105% with a solution of deionized water and 26 g/L sea salt to form a stable slurry (i.e., no particles segregated). We accounted for the residual salt in the Nankai silty claystone assuming an in situ salt content of 35 g/L and in situ water content of 27% based on moisture and density (MAD) measurements averaged over the depth range of 580.4–865.9 mbsf. For the Nankai silty claystone–silica mixtures, these values for water and salt content only applied to the claystone fraction. The silica fraction was just moistened with deionized water at a water content of 33% to avoid attracting any of the water that was reserved for the claystone. Then the moistened silica was mixed in with the Nankai silty claystone.

After homogenizing the slurry with a spatula for at least 20 min, the slurry was de-aired using a vacuum pump to eliminate any air bubbles. Then we poured the slurry into a consolidometer with an inside diameter of 6.9 cm. Below and on top of the slurry were a porous stone and filter paper allowing pore fluids to drain in both directions but preventing fines from being washed out. The initial height of all samples varied between 11.5 and 13.6 cm.

We incrementally loaded the slurry, doubling the mass on the slurry every time, to a maximum vertical effective stress of 100 kPa. A linear position transducer, which at the time of measurement could not be mounted onto the system until the beginning of the fifth stress increment due to weight restrictions, monitored vertical displacement throughout the resedimentation experiment. The last weight increment was left on the slurry until the sediment reached secondary consolidation.

After unloading the sample to an overconsolidation ratio (OCR) (maximum past effective stress of 100 kPa divided by the current effective stress of 25 kPa) of 4, we carefully and slowly extruded the samples. The final height of all samples varied between 8.2 and 9.2 cm. A small slice of material was cut off for determination of void ratio at the end of the resedimentation experiment using the wet and dry mass technique. This void ratio value was used, along with the uniaxial deformation data, to back out void ratios at any vertical effective stress assuming that the axial strain is equal to the volumetric strain. The remaining samples each yielded two specimens for consolidation testing, of which only one was tested and shown here.

Index properties

We measured the water content by oven-drying the sample at 105°C. Water content is calculated by taking the difference in the mass of the sample before and after oven-drying and dividing this difference by the oven-dried mass. We did not correct for salt content in the pore water when calculating water content. We reported the water content of the test specimen (w_n) for each experiment (Table T3). We also provided the initial void ratio (e_i) for each specimen (Table T3). Void ratio (e) is defined as the volume of voids divided by the volume of solids and is nonlinearly related to porosity (n):

From the water content and initial void ratio we calculated the initial saturation (S_i) for each specimen (Table T3):

Specific gravity (G_s) of the Nankai silty claystone was derived from MAD measurements during Expedition 322 on board the D/V Chikyu (Underwood et al., 2009). Averaging MAD measurements over the sampled depth range (586.8–774.69 mbsf) yielded a grain density of 2680 kg/m³ for the Nankai silty claystone. For the other five sediment mixtures, we linearly interpolated between both end-members assuming a grain density of 2650 kg/m³ for the silt-size silica.

Constant rate of strain consolidation testing

After resedimenting the six sediment mixtures composed of varying proportions of Nankai silty claystone from Site C0011 and silt-size silica, we conducted CRS consolidation tests according to ASTM D4186-06 guidelines (ASTM International, 2006). We used a 10,000 lb load capacity load frame, a software-controlled pump, and pressure transducers rated to 300 psi. All components were manufactured by GeoTac. The temperature in the UT laboratory was controlled to a constant 23.9°C (±0.5°C). The specimens were trimmed into a steel ring with a diameter of 4.99 cm using a trimming jig, wire saw, and sharp-edge spatula. Once the specimen was in the ring, a wire-saw, razor blade, and recess tool were used to smooth the top and bottom of the specimen and to ensure consistent specimen dimensions. The initial height (H₀) was 1.73 cm. We flushed all lines with deionized water and applied a constant backpressure of 386 kPa for at least 12 h to ensure full saturation. The interaction between deionized water in the chamber and salt water in the specimen with a high amount of swelling clays most likely caused the low saturations (Table T3).

The specimens, laterally confined in the steel ring, were consolidated at a constant rate of strain (i.e., uniaxial strain). The strain rate (dε/dt) for all six specimens varied between 0.2%/h and 1.05%/h, increasing with increasing silica content. These strain rates were not adjusted during the test, yet they ensured pore pressure ratios (ratio of excess base pore pressure to total axial stress) to be smaller than 0.1. The top of the specimen was open to the cell pressure (u_c), whereas the bottom was undrained. During the consolidation test we continuously monitored specimen height (H), total axial stress (σ_v), and base pore pressure (u). Experiments were run to a maximum vertical effective stress of 21 MPa, where stresses were held for 6 h to allow excess pore pressure to dissipate. Specimens were then unloaded to an OCR of 4.

We computed axial strain (ε), base excess pore pressure (Δu), average vertical effective stress (σ′_v), hydraulic conductivity (K), intrinsic permeability (k), coefficient of volume compressibility (m_v), and coefficient of consolidation (C_v) as follows in accordance with ASTM Standard D4186-06 (ASTM International, 2006):

and

where displacements were measured with linear position transducers. Variable definitions for all equations are in Table T4. In Figures F6–F11 we report strain as a percentage.

The calculation of intrinsic permeability from hydraulic conductivity requires knowledge of fluid properties. We used a viscosity of 0.001002 Pa·s for a temperature of 20°C and a unit weight of water of 10042.368 N/m³, assuming a fluid density of 1024 kg/m³ and an acceleration due to gravity of 9.81 m/s². We computed compression index (C_c), which is the slope of the virgin consolidation line, from the change in specimen void ratio (Δe) with vertical effective stress (σ′_v):

Maximum preconsolidation stress (σ′_pc) of all specimens is 100 kPa, equal to the maximum vertical effective stress the slurry was preloaded to during resedimentation.

SEM imaging

SEM images, both backscattered (BSE) and secondary (SE) images, were taken on a field-emission SEM. The samples were prepared using an argon-ion beam milling technique (Loucks et al., 2009), which avoids mechanical polishing and instead produces surfaces with only minor topographic variations using accelerated argon ions. BSE and SE images were taken at three different magnifications (1,000×, 14,000×, and 60,000×) after consolidation of sediment mixtures to a maximum vertical effective stress of 21 MPa. Images represent a vertical cross section of the sample (perpendicular to bedding) and are oriented so that the vertical stress was applied from the top of the image.

Top of page   |    Previous   |    Next