Previous   |    Next

doi:10.2204/iodp.proc.304305.102.2006

Physical properties

Shipboard measurements of physical properties were used to characterize recovered rocks for correlating cored material with downhole logging data and interpreting broader scale geophysical and geological data.

After recovery, the cores were allowed to come to approximately room temperature (22°–25°C), and then MS, noncontact electrical resistivity (NCR), gamma ray attenuation (GRA) bulk density, and natural gamma radiation (NGR) were measured in a series of nondestructive tests on whole cores in split liners in the MST. The bulk of Expedition 304 and 305 cores were collected with the RCB. RCB cores of hard rock are rarely full diameter, so GRA density, MS, and NGR, being rock volume dependent, are underestimated. No corrections have been applied to correct for volume variations. P-wave velocity (using the P-wave logger) was not measured on the MST, which requires full-diameter core and adequate coupling to the liner for velocity to be measured effectively. Thermal conductivity was measured nondestructively on the archive half of split core pieces. Measurements of P-wave velocity, bulk and grain density, and porosity were made mostly on 1 inch diameter minicores (Expedition 304) and ~9 cm³ cubes (Expedition 305) cut from the working half of the split core. When appropriate, these samples were also used for paleomagnetic measurements. The samples chosen for physical property measurements were generally representative of the main lithologic units and avoided veins, large phenocrysts, and so on. However, some end-member lithologies were selected for measurement. All the physical property methods except for electrical resistivity measurement are described by Blum (1997).

Nondestructive sample experiments

NCR, MS, and NGR measurements were made on the MST. At the start of Expedition 304, it was thought that resistivity and susceptibility could not be measured simultaneously, so they were measured on separate runs for Hole U1309B and the upper 130 m of Hole U1309D. Below 130 m in Hole U1309D and during Expedition 304, NCR and MS were measured concurrently. During Expedition 305, NCR, MS, NGR, and GRA were all measured. The MST limitation was actually found to be a function of the number of instruments (a maximum of four) that can be connected to the Integrated Measurements System data acquisition program at any one time. It should be noted that the calibrations assumed for the logged MST measurements of MS, GRA, and NGR assume normal full-diameter RCB core, so although the use of nonstandard sample geometries for other core types will give useful relative data, these will not be absolute measurements and will need to be corrected prior to comparison with measurements from other core types and laboratories. At present, NCR data are logged as instrument output in volts but are stored in the database in millivolts. Similarly, MS data are stored in instrument units (IU). Although major voids were skipped, it was impractical to avoid measuring small or partly broken rock pieces. In these cases, the measurements were not properly calibrated and should be ignored. Although it is appreciated that GRA density has little quantitative value during hard rock drilling expeditions, it nevertheless is a useful guide to defining density trend. GRA density also has a practical application, in that rapid decreases in GRA density can be used to map the location of breaks along the core. MST data were sampled at discrete intervals, with sampling intervals and NGR count times chosen to optimize the resolution of the data in the time available to run each section through the device (Table T11).

Electrical resistivity

Resistivity was measured using a noncontact inductive instrument, purpose-built for the MST by Geotek. The instrument transmits a high-frequency electromagnetic signal from one coil near the sample. This induces currents in a conducting sample, which in turn produces a secondary electromagnetic field that is detected by a receiver coil, also near the sample. A matched pair of transmitter and receiver coils points away from the sample and provides a null reference. The instrument measures the difference in signal received at the sample and reference coils. This difference in signal is a function of the resistivity of the sample. The instrument is rated to measure resistivity in the range of 0.1–10 Ω·m. Apparently useful measurements to >10⁵ Ω·m were obtained, but these high values are poorly calibrated. The indication of a variation in resistivity is reliable in these cases, but the measured magnitude of the change cannot be used for interpretation.

The Geotek instrument, originally fixed to the bench with Velcro, showed improved accuracy after a more permanent attachment with bolts. The instrument’s automatic zeroing was checked at the beginning of each run.

Calibration of the NCR at the beginning of Expedition 304 was conducted by measuring 13–15 cm lengths of core liner filled with several different but known concentrations of NaCl solution in water, ranging from 0.35 ppt (0.006 mol/L) to 35 ppt (0.599 mol/L). The resistivities of these standards were obtained from data provided by Geotek (Figs. F13, F14). The resulting calibration formula is

R = 0.0638 V^–1.1806,

where R is resistivity in ohm-meters and V is instrument output in volts. This formula was used for both Expeditions 304 and 305. Note that, at present, only the raw voltage values (mV) are logged in the IODP database, although for this report, the NCR data are presented in SI units based on the formula shown above. It is sometimes more intuitive to consider the reciprocal of resistivity, which is conductivity (NCC), measured in Siemens per meter (S/m).

The manufacturer, Geotek, estimates the spatial resolution of the NCR as ~2 cm. However, note that Figure F13 shows that the sensor has to be between 4 and 8 cm from the end of the sample before the full resistivity is measured; therefore, resistivity measured on pieces shorter than ~10 cm will be overestimated. NCR was measured every 2 cm along the core.

Except in cases where a rock contains large quantities of conducting minerals (mostly metal oxides and sulfides), electrical resistivity is determined by the amount and conductivity of the pore fluids in the rock, the geometry of the pore space, and the degree of alteration (e.g., Ildefonse and Pezard, 2001). As such, the degree of saturation with seawater is critical. Normally, resistivity was measured on the MST ~2 h after the core came on board (the delay being the time required for core curation), and in this time, the core can lose a significant proportion of its pore water. This dependency was investigated during drilling Hole U1309D by measuring one interval (Core 304-U1309D-38R, second uncurated liner; eventually Section 2 and part of Section 3) ~15 min after coming on board and again ~2.5 h later, after curation. The results (Fig. F15) show broadly the same trends but with differences in signal level that are summarized in Table T12. The nonlinearity of the relation between volts and resistivity and the fact that a few measurements in the 2.7 h run are quite low lead to a large standard deviation in resistivity. To reduce the large resistivity standard deviation, the test was repeated on two long pieces of core sections (Core 304-U1309D-36R-1 [Pieces 2 and 3]) from Hole U1309D lithologic Unit 88, oxide-bearing gabbros, which were relatively conductive. The working and archive halves were put together (and held in place by elastic bands) and then were remeasured dry and after saturation in seawater for periods from 2.5 to 17.7 h. The results were inconclusive (Fig. F16), and, therefore, the absolute values of conductivity should be considered “apparent” and not a direct measure of the rock properties.

Magnetic susceptibility

Magnetic susceptibility, k, is a dimensionless measure of the degree to which material can be magnetized in an external magnetic field:

k = M/H,

where M is the magnetization induced in the material by an external field of strength H. MS is sensitive to variations in the type and concentration of magnetic grains in rocks and is therefore an indicator of compositional variations.

The MST includes a Bartington susceptibility meter (model MS2C) with an 8 cm internal diameter loop sensor, which corresponds to a coil diameter of 88 mm, and operates at a frequency of 0.565 kHz with an AF intensity of 80 A/m (Bartington Instruments, Ltd., 1995). The instrument output is in volume-specific IU. For a core diameter of 66 mm and a coil diameter of 88 mm, the correction factor is 1.48 [transforming volume-specific IU to absolute SI units (× 10^–5)] (Blum, 1997). Note that because RCB and other coring procedures yield core with diameters <66 mm, the IODP database stores MS in IU.

The instrument is automatically zeroed before the sample is introduced, and a control measurement is made at the end of each run to enable correction for an assumed linear drift. The instrument can be calibrated using a standard sample of known susceptibility (e.g., a sample of known magnetite content) (Fig. F17). During Expeditions 304 and 305, RCB cores were sampled at 2.0 cm intervals. Measurement precision is 2 × 10^–6 SI, and along-core resolution is ~4 cm (Fig. F17). Note that susceptibility measured on pieces shorter than ~8 cm is underestimated.

On the Bartington MS2C sensor, all readings >10,000 IU lose the most significant digit, so the signal appears to fall discontinuously to a low value or “wrap around.” As a result, intervals where k values appear to approach 0.1 and then fall rapidly should be examined and used with care. For example, Section 304-U1309D-36R-1 in Unit 88 (oxide gabbros) shows a noisy signal, rapidly varying from very low to full range (Fig. F18). However, application of an arbitrary additive constant of either 10,000 IU or 20,000 IU (one or two times full range) makes the susceptibility signal much less noisy and better correlated with the electrical conductivity signal. In principle, it is not possible to determine a priori the additive factor to reconstruct the MS reliably. Although Figure F18 summarizes the MS problem and offers a possible solution, no attempt was made to routinely correct the MS data shown in this report or in the IODP database.

As stated above, for sufficiently small rock pieces, the measured susceptibility amplitude will be inaccurate and should be ignored.

Another instrument, the MS2E1 point sensor, is also available for measuring susceptibility on split core (on the archive MST [AMST]) (Blum, 1997). This apparently has the same characteristics as the Bartington MS2F (Bartington Instruments, Ltd., 1995) but has almost 1 order of magnitude better spatial resolution (<1 cm) than the MS2C loop sensor (4 cm) on the MST. The point susceptibility track automatically scans the core profile with a laser to sense voids, but simply observing the operation of the device indicates that, often, core spacers are measured, so care should be taken in selecting AMST data for use or comparison with the MS data. Samples were measured on the point sensor every 2 cm along the core.

MS was measured both on the MST (MS2C sensor) and on the AMST (MS2E1 point sensor). The results from the two sensors are moderately well correlated, but the MS2E1 produced values ~70% lower than those of the MS2C (Fig. F19), confirming that the approximate volume correction for the MST is often valid for hard-rock core. MS was also compared with the MS of discrete samples measured using the MS2E1 point sensor with values measured on the Kappabridge susceptibility meter in the paleomagnetism laboratory (Fig. F20). These show a good correlation with a slope of 0.57. Note that, according to the manufacturer, the MS2E1 sensor will read ~0.5 times the true susceptibility when placed against a flat surface, or 1.0 times the susceptibility “when buried up to the shoulder” (Bartington Instruments, Ltd., 1995).

For this report and in the VCDs (Fig. F2), only MST data from the MS2C sensor are displayed.

Natural gamma radiation

NGR was logged on the MST using a 30 s counting period at 10 cm intervals. The NGR sensor was built by ODP in 1993 and uses four NaI scintillation counters arranged at 90° to each other in a plane normal to the core axis. Activity was measured in total counts per second with no attempt to determine energy spectra. At background level, a count at 30 cps measured over a 30 s period gives a statistical error of ~3%. The device was calibrated for zero background by measuring a core liner filled with distilled water. Initial software problems with NGR data logging (a Y2K program bug that had only just manifested itself) resulted in no NGR measurements being made during Expedition 304 on Hole U1309B or Cores 304-U1309D-1R to 22R. However, if Expedition 305 is an indication, the NGR measurements never significantly exceeded background levels.

Thermal conductivity

Thermal conductivity was measured on the archive half of the split core. This was achieved by transient heating of the sample with a known heating power generated from a source with finite radius and assumed infinite length and infinite thermal conductivity (in practice, a 7 cm long needle). The needle probe method was used in half-space mode (Vacquier, 1985) to measure thermal conductivity in pieces of the split core taken from the archive half. The increase in temperature with time was measured using the TK04 system described by Blum (1997), and the best-fit parameters to this curve are automatically determined by the instrument, leading to an estimate of the thermal conductivity. In theory, the changing temperature should be logged for an infinite time. In practice, by fitting different portions of the curve, a series of estimates of the equilibration time is obtained together with the corresponding thermal conductivity. The minimum equilibration time and maximum standard deviation for acceptable curve fits are set to 10,000 s and 0.003 W/(m·K). Four separate runs are made on each sample, and the mean is logged. The standard deviations between the four measurements were typically ~0.1 W/(m·K). The instrument performs a self-test at the beginning of each measurement and does not need calibration. Precision has been estimated as 2%, with an accuracy of 5% (Blum, 1997).

Samples were taken at irregularly spaced intervals (typically ~1–3 samples per core, or average intervals of 1.5–3 m), depending on the availability of pieces long enough to be measured without edge effects (more than ~10 cm) and on the degree of lithologic variability. Samples needed to be quite smooth to ensure adequate contact with the heating needle; in practice, this usually meant removing any visible saw marks, and this was achieved by grinding and polishing the rock using 120 gauge silicon carbide powder. Samples were allowed to equilibrate to room temperature for at least 4 h, and then they and the sensor needle were equilibrated together in a seawater bath (enclosed within a cooler during Expedition 305) for at least 15 min prior to measurement. Isolation of the samples and sensor needle in the cooler eliminated the effect of rapid but small temperature changes introduced by the laboratory air conditioning, opening of nearby bulkhead doors, and people walking past the experiment. The instrument measures the thermal drift and will not begin a heating run until sufficient thermal equilibrium is attained. Measurements were made at room pressure and temperature and were not corrected for in situ conditions.

Thermal conductivity is an intrinsic tensor material property that depends on porosity, density, mineral composition, and fabric. Most known single-crystal thermal diffusivities are anisotropic (e.g., Kobayashi, 1974; Tommasi et al., 2002). Therefore, thermal conductivity might be expected to be anisotropic if the samples are monominerallic or if the sample is dominated by a fabric. To test for thermal conductivity anisotropy, Hole U1309B archive pieces were analyzed by taking three measurements on the cut face: one (k₀) parallel to the core and two (k₁ and k₂) at angles of approximately ±35° to the core axis (Fig. F21), these being the largest angles that can be achieved without encountering serious edge effects. An estimate of the anisotropy of thermal conductivity is then defined as

[max(k₀, k₁, k₂) – min(k₀, k₁, k₂)]/mean(k₀, k₁, k₂),

irrespective of which component was maximum. The same definition was used during Leg 209 (Shipboard Scientific Party, 2004). Note that this is only an apparent anisotropy, as it is not based on orthogonal measurements.

Each of the measurements made at 0° and ±35° to the core axis consisted of four measurements (as described above) that were averaged to obtain a single data point. The standard deviation for these four measurements, averaged over all the Hole U1309B data points, is 0.08–0.09 W/(m·K), and only two data points (both in apparently isotropic basalts) had individual standard deviations >0.12 W/(m·K). The maximum difference between the 0°, +35°, and –35° measurements, averaged over all the Hole U1309B data points, is 0.1 W/(m·K). It was concluded that, during Expedition 304, it was not possible to detect any significant anisotropy of thermal conductivity within the precision of the measurements. Thus, for Hole U1309D, measurements were made with the needle probe parallel to the core axis. Exceptions were made when the sample was too short or when it was necessary to avoid features such as veins. Rarely, when a strong lineation or foliation was present, measurements would be made as nearly parallel and perpendicular to that direction as possible.

Thermal conductivity has also been measured on gabbros and peridotites recovered from Legs 118 (Robinson, Von Herzen, et al., 1989), 147 (Gillis, Mevel, Allan, et al., 1993), 153 (Cannat, Karson, Miller, et al., 1995), 176 (Dick, Natland, Miller, et al., 1999), and 209 (Kelemen, Kikawa, Miller, et al., 2004). Except for Legs 176 and 209, these data were acquired with the Thermcon-85 system, which was replaced on board the JOIDES Resolution by the TK04 system in 1996 (ODP Leg 168). The latter probably gives more accurate and more consistent results, as the new measurement technique is less user-dependent than the older one (Blum, 1997).

Discrete sample experiments

Minicores (Expedition 304) and cubes (Expedition 305) were cut from the working halves of split cores. As with thermal conductivity measurements, material was taken at irregular intervals in an attempt to sample variations in the characteristic or representative core lithology. Their precise positions depended on the suitability of core, attempting to avoid extremely brittle zones. The average sampling rate was ~1–2 samples per core. After cutting, the samples were saturated for 24 h under a vacuum of –100 kPa. The saturated samples were then removed and wet weighed, and velocities were measured. The wet volumes of the cube samples were estimated using a caliper to measure the length of each side to 0.01 cm. Using a caliper to define wet volume for Expedition 305 cubes translated to an error of 0.018 ± 0.001 g/cm³. Samples were then placed into a dry oven at 105° ± 5°C for another 24 h period. After removal from the oven, samples were placed in a desiccator and allowed to cool to room temperature. Finally, dry weight and dry volume were measured.

P-wave velocity

The P-wave sensor (PWS3) (Hamilton Frame method) was used to measure velocities in discrete samples of the materials recovered during Expeditions 304 and 305. For Expedition 304, samples were usually 1 inch diameter minicores from the working half of the split core, drilled along the x-axis of the core normal to the split face, although occasionally, irregularly shaped pieces were used where a minicore could not be taken. For Expedition 305, cubes cut from the working half of the split were used for horizontal (x- and y-direction) and vertical (z-direction) P-wave velocity, porosity, and density measurements.

The PWS3 is a modified and updated version of the classic Hamilton Frame velocimeter, designed with one transducer fixed and the other mounted on an adjustable screw. The PWS3 is mounted vertically to measure velocities perpendicular to the core axis by placing the sample on the lower transducer and bringing the upper transducer into direct contact with the upper surface. Transducer separation was measured by a digital caliper attached to the transducers. At the beginning of Expedition 304, the lower transducer head was not adequately attached to the bottom mounting, so the mount was reinforced following the completion of Hole U1309B. Hole U1309B samples were remeasured and showed an average increase in velocity of 0.2 ± 0.2 km/s. The new values are used in this report and included in the database.

Traveltime was determined automatically using shipboard software that picked the time when the observed waveform exceeded a preset proportion of the maximum background amplitude. However, the picked time was visually seen to be later than would normally be picked by hand (Fig. F22). The arrival time was initially picked manually, but this approach was eventually abandoned for the automatic method to maintain consistency between operators. The automatic method was used for all the Expedition 304 and 305 velocity measurements reported and entered into the IODP database.

For Expedition 304, to ensure that the measured sample length corresponded accurately to the path traveled by the acoustic pulse, the end of each minicore farthest from the split face was first ground flat and parallel to the opposite face using 120 gauge silicon carbide grit with the aid of a purpose-made handheld frame. Early measurements made without this step showed poor accuracy and repeatability. When the sample was placed between the PWS3 transducers, distilled water was applied between the top and bottom of the sample and transducer heads to improve acoustic coupling (i.e., the impedance match) between the transducer and the sample. For Expedition 305 cube samples, no polishing or silicon grit was used but deionized water was used to achieve a better coupling with the transducers.

The instrument was calibrated in port and at various times throughout the expedition by measuring traveltimes over samples of different thicknesses and known velocity and determining the measured velocity by linear regression. The estimated precision is ~0.1 km/s.

P-wave anisotropy between the average horizontal and vertical velocities and horizontal anisotropy were calculated using

[mean(V_x, V_y) – V_z)]/mean(V_x, V_y , V_z)

and

(V_x – V_y)/mean(V_x, V_y),

respectively, where x, y, and z are the standard core coordinate axes, V_x and V_y are the transverse core velocities, and V_z is the longitudinal core velocity. During Expedition 304, it proved difficult either to reliably measure velocities along the minicore axis or to accurately cut cubes with sufficiently parallel faces. The parallel-blade rock saw was serviced and blades were replaced for successful cube preparation during Expedition 305. The PWS3 was used to measure each direction (x, y, and z) of the cubes, which were marked during sampling with an arrow pointing upcore in the z-direction on the x-direction face.

Porosity and density

The samples used for velocity measurements were also used to estimate bulk density, grain density, and porosity from the wet weights, dry weights, and dry volumes of the samples. The density and porosity values were determined automatically by IODP programs that convert the mass and volume measurements using the method described below. Results were then uploaded to the database.

Sample mass was determined to a precision of ±0.001 g using two Scientech electronic balances. The balances are equipped with a computer averaging system that compensates for the motion of the ship. The sample mass on one balance is counterbalanced by a known mass on the adjacent balance. The balances were calibrated in port prior to each expedition by weighing known masses. Continuing problems with the weight measurement forced recalibration of the balances to be done at sea.

The volumes of chips, minicores, and cubes were determined using a five-cell Quantachrome helium-displacement pycnometer with a nominal precision of ±0.01 cm³. Calibration was maintained by including a standard reference sphere in one of the operating cells for each run and cycling it sequentially between the cells for successive runs. The cell volumes were recalibrated if the measured volume of the standard was not within 0.02 cm³ of the known volume of the standard. Early in Expedition 304, it was found that Cell 4 did not operate correctly, and all subsequent runs were made without this cell. During Expedition 305, due to technical difficulties, the computer-compatible pycnometer had to be replaced by an older, manual pycnometer calibrated for use with medium-sized cells (as opposed to the beaker-compatible cells used for Expedition 304).

Wet weight was determined on samples that had been saturated in seawater under vacuum for 24 h. Samples were lightly patted dry on paper towels to remove excess water adhering to the surface before weighing. Dry weight and pycnometer volume measurements were then made after the samples had been oven dried for 24 h and allowed to cool in a desiccator. A potential problem with this drying temperature is that, in addition to interstitial water, chemically bound water in some clay minerals can be lost, leading to porosity errors up to 20% (Blum, 1997). This is not considered to be a problem for the gabbros, troctolites, and diabases recovered during Expeditions 304 and 305.

Water content

Water content, as a fraction of total mass or as a ratio of water mass to solid mass, is determined by standard methods of the American Society for Testing and Materials designation 2216 (ASTM, 1989). Sample saturation in seawater was done under vacuum for a 24 h period. The total water-saturated mass (M_b) and dry mass (M_d) are measured using the electronic balance as described above, and the difference is the uncorrected water mass. Measured wet and dry masses are corrected for salt assuming a pore water salinity (s) of 0.035 (Boyce, 1976). The water contents expressed as a percentage of the wet mass or the dry mass (W_b and W_s, respectively) are given by

W_b (%) = 100[(M_b – M_d)/([1 – s]M_b)]

and

W_s (%) = 100[(M_b – M_d)/(M_d – sM_b)].

Bulk density

Bulk density (b) is the density of the saturated sample,

b = M_b/V_b,

where M_b = the total water-saturated mass and V_b = the total sample volume. The latter is estimated either from the dimensions of the sample or from the volume of the dry sample (V_d) measured in a helium gas pycnometer and the volume of the pore fluid (V_pw):

V_b = V_d + V_pw .

Grain density

Grain density (g) is determined from the dry mass and dry volume measurements. Both mass and volume must be corrected for the salt content of the pore fluid:

g = (M_d – M_salt)/(V_d – M_salt/salt),

where M_d = the dry mass of the sample (in grams), salt = the density of salt (2.257 g/cm), and

M_salt = s(M_b – M_d)/(1 – s)

is the mass of salt in the pore fluid.

M_pw is the salt-corrected mass of the seawater:

M_pw = (M_b – M_d)/(1 – s),

and the density of the pore fluid, which is assumed to be seawater, is pw = 1.024 g/cm.

Then the volume of pore water is

V_pw = M_pw/pw = (M_b – M_d)/(1 – s)pw.

Alternatively, the grain density is calculated from the wet and dry masses (M_b and M_d) and the sample volume (V_b) is calculated from the measured dimensions of a minicore or cube sample:

g = (M_d – M_salt)/[(1 – s)pw/(M_b – M_d)].

Porosity

Porosity (ϕ) is the ratio of pore water volume to total volume and can be calculated from fluid density, grain density, and bulk density of the material:

ϕ = 100[(g – b)/(g – w)],

where

• g = the grain density,
• b = the bulk density, and
• w = the fluid density.

Dry density

The dry density (d) is the ratio of the dry mass (M_d) to the total volume (V_b). The dry density is calculated from the corrected water content (W_d) and porosity (ϕ):

d = (ϕ/W_d)w.

Top of page   |    Previous   |    Next