IODP Proceedings    Volume contents     Search


Laboratory testing methodology

Sample handling and preparation

All three standard coring systems, the advanced piston corer (APC), the extended core barrel (XCB), and the rotary core barrel (RCB), were used during Expedition 317. The whole-round samples used in this study were recovered using the APC and XCB coring systems only.

Because it was not possible to locate the unconformities within the cores before completion of shipboard measurements, the samples for uniaxial consolidation tests were taken according to the following scheme. The first hole (Hole A) drilled at each site was dedicated to whole-round sampling for microbiology, geochemistry, and geotechnical studies, and we obtained 10 cm long whole-round samples for uniaxial consolidation tests from the lower end of each core. From Holes B and C at each site we collected 6 cm long whole-round samples from the lower ends of approximately every fifth core. Core catchers, cores from Holes B and C that duplicated the section cored in Hole A, cores filled with pure sand, obviously disturbed cores, or cores with poor or very poor recovery were not sampled.

To avoid any additional stress on the samples or fracturing prior to the experiments and to maintain natural saturation, the whole-round samples were not extruded but cut off while still in the core liners, capped, sealed in wax, and shrink-wrapped. The samples were kept in refrigerated storage at ~4°C both on the ship and on shore.

For the experiments, each sample was carefully removed from its wax-sealed liner and subsampled with a sharp cutter and a cutting shoe. Remaining material was capped, sealed in wax, and stored under refrigeration again for potential additional measurements.

Sample descriptions

All samples measured were visually undisturbed and showed no obvious deformation or inhomogeneity. The grain size of the samples ranges from clay to silty clay. Some samples contained few shell fragments ranging in size from the sand to coarse sand fraction. One exception, Sample 317-U1351B-104X-3, 100–106 cm (912.8–912.86 meters below seafloor [mbsf]), contained mainly silty and fine-grained sand and many shell fragments. As a result, this sample was too fragile; it crumbled into pieces during sample preparation and was not tested.

Index properties

Water content (wc) and bulk density (ρb) were measured on the test specimens during the onshore consolidation tests by oven-drying the samples. Water content was calculated by taking the difference in the weight of the sample before and after oven-drying and dividing this difference by the oven-dried weight. Bulk density was calculated by dividing the wet weight of the sample by the volume of the specimen. We compared the bulk density measured on our specimens with the shipboard measurements of moisture and density (MAD). The laboratory-derived wet bulk density is very scattered and underestimates density relative to the MAD wet bulk density by a factor of ~1.08 (Fig. F2). For Holes U1352A and U1354A, no shipboard MAD data are available, so we used the laboratory-derived bulk density only.

Uniaxial consolidation testing

The uniaxial consolidation tests were performed in the geotechnical laboratory of MARUM–Center for Marine Environmental Sciences (Bremen, Germany). We conducted incremental loading consolidation tests in a fixed-ring oedometer system following the general standard for unaxial consolidation tests specified by the Deutsches Institut für Normung (German Institute for Standardization) (DIN, 1999). The nomenclature and symbols used in this study are summarized in Table T1.

Specimens were prepared from the centers of whole-round samples parallel to the longitudinal axis of the core. To avoid failure of the grain structure, which would have a negative effect on preconsolidation stress, each sample was trimmed carefully by hand with a cutter and then placed in a tightly fitting stainless steel specimen ring. The specimen ring was used to maintain a condition of zero lateral strain. The ring has an inner diameter (d) of 5.05 cm (area [A] = 20 cm2) and a height (hi) of 1.48 cm. A smaller ring size with d of 3.57 cm (A = 10 cm2) and hi of 14.8 mm was used if it was not possible to prepare intact samples for the larger ring size. This was the case for five samples (Table T2). However, the unfavorable diameter/height ratio of the small ring size may result in unwanted edge effects of friction resistance. Immediately after trimming the sample into the specimen ring, the wet weight of each sample was determined.

Loading of the specimen was applied through a mechanical lever arm capable of a maximum vertical stress of 18,000 kPa for an area of 20 cm2. Nine incremental loading steps were needed for samples from the upper 100 m and 12–15 steps were needed for samples from core depths below 100 mbsf. At the beginning of each experiment, the specimen was saturated with demineralized water in the oedometer cell for 24 h with applied initial vertical stress of 6.13 kPa. The use of demineralized water, which does not correspond chemically to the natural pore water of the samples, did not obviously affect the displacement behavior of the sample, and no change in height of the sample was observed under normal pressure conditions.

For each loading phase, the initial applied vertical stress (σ′vi) was doubled to a maximum vertical stress (σ′vm) of up to 4,905 kPa for samples from core depths above 100 mbsf or 34,320 kPa for samples from core depths below 100 mbsf. We allowed 24 h for each loading step until the applied vertical stress was nearly equal to the vertical effective stress (σ′v). The highest loading rates were applied on specimens from the deepest intervals to ensure that the effective stress applied to the specimen would exceed the preconsolidation stress. For the unloading phase, we reduced the applied vertical stress by ¼ every 24 h. The associated change in height (h) was measured continuously with a linear variable differential transformer mounted on the top of the consolidation cell and digitally recorded through an analog-to-digital converter.

The change in void ratio (e) during the test was calculated following the approach presented by (Blum, 1997):

e = (Vb × 2.65 – wd)/wd,


  • Vb = bulk volume of the specimen (in cubic centimeters) given by the height of the sample at the end of each loading phase and the specimen ring area,

  • 2.65 = assumed grain density (in g/cm3), and

  • wd = dry weight of the sample after the test by oven-drying for 24 h at 60°C.

We computed the compression index (Cc) as a function of the change in the void ratio in relation to the vertical effective stress during normal consolidation (Craig, 2004):

Cc = –(en1en)/log(σ′v n/σ′v n1).

This relationship is also shown graphically as the slope of the virgin compression curves (blue line) in the result plots of the uniaxial consolidation tests (Figs. F3–F38; Table T2; see OEDOMETER in “Supplementary material”).

Preconsolidation stress (σpc) is the apparent maximum vertical effective stress the sediment experienced. σpc was determined by the graphical method of Casagrande (1936). This method requires that the transition from the reloading curve to the virgin compression curve is well defined on the semilogarithmic plot of void ratio versus vertical effective stress. σpc (green dot in Figs. F3–F38) corresponds to the abscissa of the point of intersection between the backward projected virgin compression line (blue) and the bisecting line (black solid) between the horizontal line (black dashed) and the tangent (black dotted) at the point of maximum curvature of the consolidation curve (red). The Casagrande method potentially underestimates σpc, and preconsolidations >800 kPa are potentially untrustworthy (Sauer et al., 1993).

The constrained modulus (Es) is a specific value for computation of the subsidence behavior of a soil and defines the soil resistance against deformation:

Es = (Δσ′v/Δε) × (1 – ε),

where ε is the strain rate, the change in specimen thickness for each load step in relation to the initial specimen thickness (Δh/hi).

The consolidation state of the sediments was estimated from the overconsolidation ratio (OCR), which is defined as the ratio of the preconsolidation stress and the vertical hydrostatic effective stress (σ′vh) in the following equation:

OCR = σpc/σ′vh,

and σ′vh was computed by

σ′vh = g × zs × (ρb – ρpw),


  • g = acceleration of gravity (9.81 m/s2),

  • zs = midpoint of the sample depth interval (mbsf),

  • ρb = average bulk density from shipboard MAD measurements close to the relating depth interval or from uniaxial consolidation tests (g/cm3), and

  • ρpw = pore water density (~1 g/cm3).

For the calculation, it is assumed that only hydrostatic pore pressure exists in the sediments. If OCR = 1, the tested sample is considered to be normally consolidated; OCR < 1 indicates underconsolidation, and OCR > 1 indicates overconsolidation.