IODP Proceedings    Volume contents     Search

Previous   |    Next

doi:10.2204/iodp.proc.330.102.2012

Physical properties

Shipboard measurements were performed to characterize the physical properties of the recovered material. The primary objective of Expedition 330 was to recover hard rock, which necessitated a modified approach compared with expeditions focusing on the recovery of soft sediment. The descriptions herein relate specifically to hard rock material; sedimentary material was treated in the same manner, except where otherwise specified.

Once recovered, whole-round cores were first allowed to thermally equilibrate to ambient room temperature (~3 h for both hard rock and sedimentary material). Core sections with continuous intervals of >8 cm were run through the Whole-Round Multisensor Logger (WRMSL) for measurement of gamma ray attenuation (GRA) bulk density and magnetic susceptibility. The WRMSL also incorporates a noncontact resistivity detector and a compressional wave (P-wave) velocity logger. The noncontact resistivity detector was not used on this expedition, and the P-wave logger was used only for sections where unconsolidated sediment filled most of the core liner because the diameter of hard rock cores is generally too small to make good contact with the core liner. Sections longer than 50 cm were measured with the Natural Gamma Radiation Logger (NGRL).

After measurements with the WRMSL and NGRL, the cores were split into archive and working halves. The archive half of the core was passed through the Section Half Multisensor Logger (SHMSL) for measurement of point magnetic susceptibility and color reflectance. The SHMSL uses a laser to create a profile of the archive half of the core, which yields information about the location of gaps and cracks between pieces of the core. This information was used to filter out data from gaps and cracks between pieces in magnetic susceptibility, color reflectance, and GRA data using a custom-designed MATLAB program suite, as described in “Data filtering.” The filtered data were then visually double-checked against images of the core section halves.

Nondestructive thermal conductivity testing was performed on the working halves of the split cores. Until failure of the thermal conductivity instrument prevented further testing (see “Thermal conductivity”), roughly one measurement was made per core, depending on core recovery and lithologic variability.

Lastly, discrete samples were taken from the working half of the core with a sampling interval of approximately 2–3 m, depending on lithologic variability. Many of these discrete samples were used for paleomagnetic measurements of alternating-field demagnetization (see “Paleomagnetism”), before being used for measurements of compressional wave velocity in three orthogonal directions and moisture and density measurements of wet bulk density, dry bulk density, grain density, water content, void ratio, and porosity. Details about each physical property measurement are given below. A comprehensive discussion of methodologies, calibrations, and calculations used in the JOIDES Resolution physical properties laboratory is presented by Blum (1997).

Whole-Round Multisensor Logger measurements

GRA-derived bulk density and magnetic susceptibility were measured nondestructively with the WRMSL. To optimize WRMSL performance, sampling intervals and measurement integration times were the same for all sensors. Sampling intervals were set at 2 cm, with an integration time of 5 s for each measurement. These sampling intervals are common denominators of the distances between the sensors installed on the WRMSL (30–50 cm), which allows sequential and simultaneous measurements. After every core, quality control and assurance were monitored by passing a single core liner filled with deionized water through the WRMSL. For Expedition 330, the primary objective was the recovery of hard rock, which is often recovered in pieces rather than intact core. Anomalously low values associated with gaps and cracks in the core were removed using a data filtering procedure (see “Data filtering”).

Gamma ray attenuation bulk density

The GRA densitometer on the WRMSL operates by passing gamma rays from a 137Cs source through a whole-round core into a 75 mm × 75 mm sodium iodide detector situated directly below the core. The gamma ray peak has a principal energy of 0.662 MeV that attenuates as it passes through the core. The attenuation of gamma rays occurs primarily by Compton scattering, which is related to the material bulk density; therefore, for a known thickness of sample, the density (ρ) is proportional to the intensity of the attenuated gamma rays and can be expressed as

ρ = ln(I/I0)/(µd),

where

  • I = measured intensity of gamma rays passing through the sample,

  • I0 = gamma ray source intensity,

  • µ = Compton attenuation coefficient, and

  • d = sample diameter.

µ and I0 are treated as constants, such that ρ can be calculated from I.

In general, WRMSL measurements are most accurate when taken on a completely filled core liner with minimal drilling disturbance; otherwise, measurements tend to underestimate true values. By default, the instrument reports measurements using the internal diameter of the core liner (66 mm) as the assumed sample diameter. This assumption is suitable for most sediment cores; however, for hard rock the core diameter is usually <58.5 mm. Following Jarrard and Kerneklian (2007), the density measurements were corrected by multiplying the density values by 66/58 = 1.138 to account for this bias.

The spatial resolution of the GRA densitometer is less than ±1 cm. Calibration of the densitometer was done using a set of aluminum cylinders encased in a core liner filled with distilled water. Recalibration was performed as needed when the deionized water quality control and assurance standard deviated significantly (more than a few percent) from 1 g/cm3.

Magnetic susceptibility

Magnetic susceptibility (K) is a dimensionless measure of the degree to which a material can be magnetized by an external magnetic field:

K = M/H,

where M is the magnetization induced in the material by an external field of strength H. Magnetic susceptibility is sensitive to the type and concentration of magnetic minerals within a material, making it useful in identifying compositional variations. Materials such as clay, possibly from alteration of igneous materials, have a magnetic susceptibility several orders of magnitude lower than magnetite and some other iron oxides that are common constituents of igneous material. Water and plastics (e.g., core liner) have a slightly negative magnetic susceptibility.

The WRMSL incorporates a Bartington Instruments MS2 meter coupled to an MS2C sensor coil. Data from this instrument require a correction factor, which depends on both the instrument and the operating frequency. For Expedition 330, the instrument operated at a frequency of 0.621 kHz, whereas the sensor assumes an operating frequency of 0.595 kHz. The data are multiplied by the appropriate factory-supplied correction factor of 1.099 before being uploaded to the LIMS database. These values, however, must be converted to dimensionless SI units by multiplying by 10–5. Measurements from the MS2C sensor are also sensitive to the diameter of the core relative to the diameter of the sensor coil and thus require an additional correction. The true magnetic susceptibility of the core is given by

K = αKrel,

with

α = 0.290(D/d)3,

where

  • K = magnetic susceptibility of the core,

  • Krel = measured magnetic susceptibility,

  • D = coil diameter (i.e., the aperture diameter plus 8 mm) (see MS2 Magnetic Susceptibility System operation manual, www.bartington.com/operation-manuals.html), and

  • d = core diameter.

For the instrument on board the JOIDES Resolution, D = 88 mm, which gives a value of α = 1.0 for hard rock cores (with d generally between 58 and 58.5 mm) and α = 0.687 for sediment cores (with d = 66 mm). However, for consistency with previous expeditions, we followed Blum (1997) and used α = 0.68 for intervals containing unconsolidated sediments, corresponding to a core diameter of 66.1–66.2 mm. Data discussed in the site reports and shown on the VCDs reflect corrected values. However, data in the LIMS database remain uncorrected, with only the factory-supplied correction applied.

The spatial resolution of the MS2C instrument is ±2 cm; therefore, magnetic susceptibility of core material that is not continuous over a 4 cm interval will be underestimated. Data collected near gaps or from pieces <4 cm in length were removed by a filtering process (see “Data filtering”). The removed data are shown as shaded points in the VCDs but are not shown in the site report figures. Quality control and assurance was performed at the start of the expedition with a factory-provided standard of a homogeneous mixture of magnetite and epoxy (to prevent oxidation of the magnetite) in a 40 cm long piece of core liner.

Natural Gamma Radiation Logger

The NGRL installed on the JOIDES Resolution was designed and built at the Texas A&M University IODP-USIO facility from 2006 to 2008. Natural gamma radiation (NGR) arises primarily as a result of the decay of 238U, 232Th, and 40K isotopes. Data generated from this instrument are used to augment geologic interpretations, including delineation of flow boundaries and alteration zoning throughout the core. The NGRL consists of 8 sodium iodide (NaI) scintillator detectors, 7 plastic scintillator detectors, 22 photomultipliers, and passive lead shielding. The NaI detectors are covered with 8 cm of lead shielding. Half of the lead shielding closest to the NaI detectors is composed of low-background lead (~3 Bq/kg internal radioactivity rate), and the outer half is composed of regular (virgin) lead (50–200 Bg/kg). Lead separators composed of ~7 cm of low-background lead are also positioned between the NaI detectors. In addition to passive lead shielding, the NGRL employs a plastic scintillator that suppresses data collection when high-energy gamma and muon components of cosmic radiation pass through it.

For reporting purposes, the counts are summed over the range of 100–3000 keV to be comparable with data collection from previous expeditions and for direct comparison with downhole logging data. Background measurements of an empty core liner were made for a total count time of 11 h when moving to new sites located at least 150 nmi from the previous site (~15 h of total transit time); this background radiation was then automatically subtracted from each new measurement. For Site U1377, technical difficulties with the NGRL software meant that background measurements could only be taken for 4.5 h instead of the standard 11 h.

NGRL measurements are taken at a spatial resolution of 10 cm for a total of 16 measurements on a standard 150 cm long section of core. Because the detector unit contains eight sensors positioned 20 cm apart, analysis of each section of core consists of two steps: the first set of eight measurements at 20 cm apart, followed by a shift of the core section by 10 cm for the second set of eight measurements. For each section, the count time was the same at each position. Sections <50 cm in length were not logged.

The quality of the NGRL data depends on the concentration of radionuclides in the sample and also on the counting time, with higher times yielding better counting statistics and more accurate measurement of the energy spectra for each detector. It may be possible during postexpedition analyses to separate high-quality energy spectra into 40K, 232Th, and 238U contributions, leading to eventual quantification of concentrations of the radionuclide daughters. During Expedition 330, count times ranged from 0.25 to 1 h for each position, resulting in total count times of 0.5–2 h per section, depending on core recovery and overall core flow. There was a concerted effort to keep count times as high as possible for future work on total abundance calculations, but ultimately times had to be reduced to keep up with exceptionally high recovery, particularly at Site U1374. For Sections 330-U1374A-42R-3 through 73R-1, count times were reduced to 60 min for sections containing potential lava flows or lobes and 30 min for all other sections. For Site U1377 cores, count times were reduced from 60 to 30 min, as needed, to ensure core was processed as quickly as possible before the end of the expedition. The counting statistics for these shorter measurement times are still of acceptable quality because the alkalic rocks of the Louisville Seamount Trail contain significantly higher K, U, and Th concentrations than mid-ocean-ridge basalt oceanic crust.

The NGRL was calibrated using a 137Cs and 60Co source, identifying the peaks at 662 keV (137Cs) and 1330 keV (60Co) using calibration materials obtained from Eckert and Ziegler Isotope Products (Valencia, California [USA]).

Section Half Multisensor Logger measurements

The SHMSL measures color reflectance and magnetic susceptibility on core section halves. The archive half of the split core is placed on the core track, above which an electronic platform moves along a track, recording the height of the split-core surface with a laser sensor. The laser establishes the location of the bottom of the section, and then the platform reverses the direction of movement, moving from bottom to top making measurements of point magnetic susceptibility and color reflectance. During Expedition 330, point magnetic susceptibility and color reflectance data were collected at 1 cm intervals, the highest possible resolution.

Color reflectance spectrometry

Color reflectance of the core was measured using a color reflectance spectrometer that takes measurements from 380 to 900 nm wavelengths at 2 nm intervals using an Ocean Optics 30 mm integrating sphere and both halogen and LED light sources, covering wavelengths from ultraviolet through visible to near-infrared. The scan of the entire wavelength range takes ~5 s per data acquisition offset. From the full reflectance spectra collected, the data are reported using the L*a*b* color system, in which L* is lightness, a* is redness (positive) versus greenness (negative), and b* is yellowness (positive) versus blueness (negative) of the rock. The color reflectance spectrometer calibrates on two spectra: pure white (reference) and pure black (dark). Color calibration was conducted approximately once every 6 h (twice per shift).

Point magnetic susceptibility

Point magnetic susceptibility was measured with a Bartington MS2 meter and an MS2E contact probe with a flat 3.8 mm × 10.5 mm rectangular sensor with a field of influence of 5 mm and an operation frequency of 2 kHz. The instrument averages three measurements from the sensor for each 1 cm offset, leading to an accuracy of ~5%. The spatial resolution of the point magnetic susceptibility instrument is ~3.8 mm, making it better than whole-round magnetic susceptibility for sections containing broken pieces <4 cm in length (the spatial resolution of whole-round magnetic susceptibility). As with whole-round magnetic susceptibility measurements, the output displayed by the point magnetic susceptibility sensor must be converted to dimensionless SI units by multiplying by 10–5. The probe is zeroed in air before each measurement location to avoid influence from the metal track. The point magnetic susceptibility meter was calibrated by the manufacturer before installation on the ship and is quality checked every ~6 h at the same time as color reflectance sensor calibration.

Thermal conductivity

Thermal conductivity is the measure of a material’s ability to conduct heat and is sensitive to the material’s porosity, composition, and microstructure. Measurements of thermal conductivity were made by transiently heating samples with an automatically determined heat flux controlled by the heating power and duration using either half-space or full-space geometry. The changes in temperature with time after initial heating were recorded using a Teka TK04 instrument system. The temperature of the superconductive probe has a linear relationship with the natural logarithm of the time after the initiation of heating (Blum, 1997).

For lithified sections, thermal conductivity was measured on the split core with the probe in half-space mode using a needle probe embedded in the surface of an epoxy block with a low thermal conductivity (Vacquier, 1985). Sample surfaces need to be smooth to ensure adequate contact with the heating needle. Silicon gel can be employed to improve sample/needle contact, but this was avoided during Expedition 330 in order to keep the samples clean.

Cores were measured at irregular intervals at an average of once per core, depending on the availability of homogeneous and relatively vein- and crack-free pieces that were long enough to be measured without edge effects. Pieces should be at least 7 cm long, longer than the instrument needle. Samples were selected from the working half and returned unaltered to the core liner upon completion of the tests. The sample and sensor needle were equilibrated together in seawater in a cooler insulated with extruded polystyrene foam for at least 1 h prior to measurement. Isolation of the sample and sensor needle eliminated the effect of rapid but small temperature changes introduced by air currents in the laboratory.

Thermal conductivity in unlithified sediment was measured once per core when deemed appropriate, depending on the degree of presumed drilling disturbance. These measurements were taken using a needle embedded in the center of the section, with the probe in full-space mode.

At the beginning of each measurement, temperatures in the samples were monitored to ensure that the background thermal drift was <0.04°C/min. After the background thermal drift was determined to be stable, the heater circuit was closed and the increase in the probe temperature was recorded. Maximum sample temperature during the measurements did not exceed 25°C for sediment or hard rock.

Thermal conductivity values were determined from the average of three repeated measurements. Individual measurements are usually within 1% of the mean for both full- and half-space measurements and thus within a total uncertainty of 5% (Blum, 1997). Measurements were made at room temperature and pressure and were not corrected for subsurface conditions. A MACOR ceramic standard with a certified thermal conductivity of 1.637 ± 0.033 W/(m·K) (TeKa, Version 4.0.2) was tested frequently as a quality assurance check to ensure that readings remained accurate.

Unfortunately, during thermal conductivity testing on core intervals from Site U1374, the hard rock thermal probe puck began to yield unreliable values, intermittently producing no value at all and then becoming completely nonfunctional. A backup thermal conductivity probe puck was then used, and it yielded one additional reasonable data point before failing as well. The probes were thoroughly dried and then used with the MACOR ceramic standard to check for accuracy. Each puck produced results for several tests but returned values significantly lower than the standard 1.637 ± 0.033 W/(m·K) value before failing completely once again. These probe malfunctions are attributed to water seeping into the electronic wiring, thus shorting out the device and corroding the wiring. This corrosion changed the resistivity of the wiring (to which these measurements are very sensitive), and thermal conductivity measurements could not be made for the remainder of the expedition.

Discrete samples

Cubic samples were cut from the working halves of split cores at an interval of approximately one sample for every two sections. These ~8 cm3 samples were selected to best represent the major stratigraphic units and any variation in lithology. To minimize core depletion, most discrete samples were used for paleomagnetic measurements of alternating-field demagnetization (see “Paleomagnetism”) before being used for compressional wave velocity and moisture and density measurements, as discussed below.

Moisture and density

Several rock properties (water content or moisture, bulk density, dry density, porosity, and void ratio) were determined through precise mass and volume determinations on discrete samples. Moisture and density data were also used for comparison with GRA-derived bulk density data from the WRMSL. The shipboard moisture and density facility on the JOIDES Resolution for hard rock coring consists of a vacuum water saturator, a dual balance system, and a pycnometer.

Sample preparation

Determination of a precise and accurate wet mass of sparsely porous material requires that the pore space of the sample be completely saturated. To do this, we placed the samples in individual plastic vials filled with seawater and used a vacuum chamber. A vacuum pump removed the air from the chamber to a pressure of ~40–50 kPa below atmospheric pressure, forcing the seawater into the samples. The samples were kept under saturation for at least 24 h. After removal from the saturator, the cubes remained in the plastic vials filled with seawater to help prevent evaporation of pore water. Next, the cube surfaces were patted dry with a paper towel and wet mass was immediately determined using the dual balance system. Compressional wave (P-wave) velocities were then measured on the wet sample. Following the velocity measurements, the samples were dried in a convection oven for at least 24 h at 105° ± 5°C. Dried samples were then cooled in a desiccator for at least 60 min before the dry mass was measured.

Dual balance system

The dual balance system was used to measure both wet (MW) and dry (MD) masses. Two analytical Mettler-Toledo XS204 balances were used to compensate for ship motion: one acting as a reference and the other for simultaneously measuring the unknown. A standard weight with similar mass to the sample (±1 g) was placed on the reference balance to increase accuracy, giving a precision of better than 0.005 g (Expedition 324 Scientists, 2010). The default setting of the balances is 300 measurements, which takes ~1.5 min. However, in rough seas and during transit we found that longer measurement times of 400–1500 counts, depending on sea state, were required to keep the error range at or below 0.001 g. In addition to electronic records, the final masses and count, if different from 300, were also recorded on data log sheets, hard copies of which are stored at IODP at Texas A&M University.

Pycnometer system

The pycnometer system measures dry sample volume (VD) with a nominal precision of ±0.04 cm3. Constant-volume gas pycnometers allow volume measurements to be made by evaluating gas pressures between two sealed chambers: a sample chamber and an expansion chamber. The sample chamber is pressurized to an initial pressure (P1), and then the gas is allowed to expand into the second chamber. The sample volume, VS, is then given by

VS = VC + [VE(PA – P2)]/(P1 – P2),

where

  • VC = known volume of sample chamber when empty,

  • VE = known volume of expansion chamber,

  • PA = ambient pressure, and

  • P2 = pressure after gas is allowed to expand into both chambers.

For Expedition 330, the pycnometer system on board the JOIDES Resolution had five working sample chambers, each with its own expansion chamber, allowing as many as four samples to be measured per run.

At the start of the expedition and midway through when the helium gas tank was changed, shipboard technicians performed a calibration using stainless steel spheres of known volume to determine the volume of the sample and expansion chambers. During both acquisition and measurement, the three data acquisition cycles were preceded by three purges of the sample chambers with research-grade (99.995% or better) helium gas heated to 28°C. These purges help remove any volatiles that could affect chamber pressure. A batch of samples consisted of four cells with unknowns and one cell with two stainless steel spheres (3.2 and 7.0 cm3), assumed to have a total volume known to within 1%. The spheres were routinely cycled through the different cells to identify any systematic error or instrument drift. The final volume for each sample was recorded on the data log sheets together with mass information.

Moisture and density calculations

Following physical measurements of wet mass (MW), dry mass (MD), and dry volume (VD), the physical properties shown in Table T9 were calculated. During Expedition 330, moisture and density Method C was used to increase the accuracy of the wet mass measurements (Blum, 1997).

Compressional wave velocity

Using the same discrete cube samples as were used for moisture and density measurements, the compressional wave (P-wave) velocity of hard rock samples was measured immediately after determination of wet mass. The system uses Panametrics-NDT Microscan delay line transducers, which transmit at 0.5 MHz. To maximize contact with the transducers, deionized water was applied to sample surfaces before the transducers were placed on the sample. Measurements used x-axis caliper-type contact probe transducers on the P-wave velocity gantry, and oriented samples were rotated manually to measure the other two axes. The three orthogonal axes form a right-handed coordinate system and are defined as follows: the x-axis is normal to the split-core surface and pointing into the working half, the y-axis is horizontal and parallel to the split-core surface, and the z-axis is downcore. The reference frame is identical to that used for paleomagnetism and structural geology analyses (see “Structural geology” and “Paleomagnetism”).

The signal received through the sample was recorded, and the peak of the initial arrival was chosen with automatic-picking software. The complete waveform was automatically stored by the system, should reanalysis be deemed necessary; however, shipboard visual checks of the picks appeared satisfactory. The distance between transducers was measured with a built-in linear variable differential transformer (LVDT).

Calibration was performed before measurements were made at least once a day using a series of acrylic cylinders of differing thicknesses and known P-wave velocities of 2750 ± 20 m/s. The system time delay determined from calibration (i.e., the delay observed using the acrylic standards) was subtracted from the picked arrival time to give the traveltime of the P-wave through the sample. The thickness of the sample (calculated by the LVDT in meters) was divided by the traveltime (in seconds) to calculate a P-wave velocity in meters per second.

Data filtering

To remove spurious data in hard rock core sections related to gaps and piece breaks, magnetic susceptibility, GRA-derived bulk density, L*, a*, b*, and point magnetic susceptibility data were filtered using a MATLAB program suite. These programs were written by Expedition 330 scientists L. Kalnins and P. Fulton and were inspired by a similar procedure written by A. Harris (Expedition 324 Scientists, 2010). This program uses sample height data collected with the laser attached to the SHMSL to identify cracks between pieces and gaps where plastic spacers were inserted between pieces. These gaps and cracks will, in general, provide anomalously low GRA-derived bulk density and whole-round and point magnetic susceptibility values, as well as irrelevant color reflectance values. The NGR whole-round data were not filtered because of the large sampling interval.

Cracks and gaps in the core were detected using a combination of the curated piece lengths and positions, a minimum laser height, and the gradient of the laser height. Starting from the curated piece divisions, the filter attempts to find a gap matching each division, defined as a region bounded by gradients in excess of 40% (the piece edges) or falling below the minimum laser height. If a gap cannot be detected (for example if the pieces are pressed against the plastic divider), the curated piece division is used. Pieces are then further divided using two criteria: first, any gaps detected above (bounded by gradients in excess of 40% or falling below the minimum laser height) that did not correspond to a curated piece division and, second, smaller cracks with a gradient of >10% and an extent of <1 cm. Limiting the extent of this second detection threshold to 1 cm prevents removal of pieces that are not horizontal in the core liner. Finally, pieces <2 cm in length were removed and the subdivisions were checked against core photographs to remove spurious piece breaks (e.g., breaks due to large vesicles, which are indistinguishable from cracks using a single laser profile). Data measured in the spaces between pieces, as well as measurements taken near piece edges, were then removed. Finally, remaining outliers were removed using an iterative implementation of the Grubbs test (Grubbs, 1969).

The gradient thresholds to test for both gaps and cracks, as well as minimum piece length, can be set by the user. The values used here were experimentally tuned to balance the number of divisions not detected with the number of spurious piece breaks introduced, based on laser profiles with a measurement every 0.1 mm (accurate to ±0.02 mm). In the filtering program, the edge effect (i.e., how far a measurement must be from a gap or crack to remain unaffected by the missing material) associated with each instrument may also be set by the user, together with the minimum length that a piece must be to provide an acceptable measurement, typically twice the edge effect. Parameter settings used to create filtered data sets for this expedition are listed in Table T10.

Data that were filtered out appear in the VCDs as shaded points, whereas data that met the criteria described above (e.g., GRA-derived bulk density measurements that were at least 1 cm away from a gap or crack and that were measured on pieces that were at least 2 cm long) are plotted as solid colored points according to the legend. Only data that met the filtering parameters are shown on the figures in the site reports. For GRA-derived bulk density and whole-round magnetic susceptibility data, the corrections to account for the average core diameter were applied even to data that were filtered out, because these corrections resulted from instrumentation design and are not related to the filtering process.

The major benefit of filtering GRA-derived bulk density, magnetic susceptibility, and color reflectance data is that actual variations can be discerned from data trends. Data that are filtered out of the GRA-derived bulk density and magnetic susceptibility data sets are underestimates of real values because of edge effects. However, caution should be used because filtering parameters have been set to remove a minimum amount of data, so some data affected by discontinuities in the core may remain.

Top of page   |    Previous   |    Next