31. Effect of Hydrostatic Pressure on the Thermal Conductivity of Basalt from Hole 504b, with Some Reassessment of the Shipboard Data

Thermal properties (diffusivity and heat capacity) at varying temperatures up to 200°C and the thermal diffusivity under hydrostatic pressures up to 1 GPa at room temperature were measured on two basaltic samples recovered from Hole 504B drilled during Ocean Drilling Program Leg 148. Rock fabrics and porosity are key factors controlling the thermal properties of these rock samples. Estimates of thermal conductivities of rocks in situ can be made on the present results. The estimated values can be compared with those obtained by shipboard measurements, enabling us to eliminate incorrect data. The shipboard measurements of rock samples containing a small amount of water-filled pores give reliable thermal conductivity values. The validity of measuring wet samples also is verified by thermal diffusivities of dried core samples, measured experimentally under hydrostatic pressures. INTRODUCTION Thermal conductivity data are available from the top to the bottom of Hole 504B. These samples penetrate through the sedimentary layer down to the Layer 2/3 (basalt/gabbro) boundary. Heat-flow values and lithospheric thickness help to constrain the formation age of the oceanic basin to as old as 80-120 Ma. Heat-flow values can be obtained by combining the subseafloor temperature gradient with the in situ thermal conductivity. The thermal conductivity usually is obtained by measuring samples on board at laboratory temperature and atmospheric pressure. Depressurization of rock samples deep from the basement increases the volume of pores in each sample, which may have a large effect on the thermal conductivity values. This study clarifies experimentally the influence of porosity on the thermal properties (diffusivity, heat capacity, and conductivity) in order to obtain intrinsic (or rock matrix) thermal conductivity by applying the formulae of Walsh and Decker (1966), by which the matrix conductivity is obtained by extrapolating values at high pressure to zero pressure conditions. If this is applicable, the experimental difficulty for measuring the conductivity of small and irregularly shaped hard rock samples will be significantly diminished. Another goal of this study was to determine whether hydrostatic pressue significantly affects the intrinsic matrix thermal conductivity. If it does, heat-flow data obtained by conventional methods must be corrected both for porosity and for the pressure dependence of matrix conductivity. Pressure measurements in the present experiment were made from atmospheric pressure to 1 GPa. METHODS Samples measured in this study were cut into cylinders 5.0 mm in diameter and 10.0 mm in length (for thermal expansion in vacuum), 'Alt, J.C., Kinoshita, H., Stokking, L.B., and Michael, P.J. (Eds.), 1996. Proc. ODP, Sci. Results, 148: College Station, TX (Ocean Drilling Program). Japan Marine Science and Technology Center, Headquarters, 2-15 NatsushimaCho, Yokosuka 237 Japan, [email protected] 3 Earthquake Research Institute, University of Tokyo, 1-1-1 Yayoi, Bunko-ku, Tokyo, 113 Japan. "Institute for Study of the Earth's Interior, Okayama University, 827 Yamada, Misasa, Tottori, 682 Japan. Faculty of General Education, Osaka Sangyo University, 3-1-1 Nakakakiuchi, Daito-shi, Osaka, 574 Japan. 10.0 mm in diameter and 20.0 mm in length (for thermal diffusion under pressure), or into disks 10.0 mm in diameter and 2.5 mm in thickness (for heat capacity and thermal diffusivity in vacuum). Thermal expansion results are not reported in this article. Both ends of the samples were polished using alumina powder of 600 mesh (the grain size of phenocrysts is far smaller than 2.5 mm.). The samples were dried in a furnace at 100°-l 10°C for a couple of days. The disk samples were used for measuring heat capacity, thermal diffusivity, density, and porosity. All measurements were made at room temperature (15°-20°C). Heat capacity and diffusivity was measured using the flash method (e.g., Watanabe, 1992) in vacuum of 10" Torr. (0.132 Pa). Density (dry-bulk density) and porosity were obtained by comparing the weights of dried and water saturated samples on a highsensitivity balance open to laboratory air. Saturation by water was made by placing the sample in a hydraulic pressure of 1 GPa, which is high enough to close pores (granite sample, Walsh and Decker, 1966). The problem is briefly referred to in the footnote of Table 1. Thermal diffusivity was measured by placing a heater wire in the center and two thermocouples 3 and 4 mm off center of the cylindrical sample. The basic idea of the present method is similar to the Angstrom method (e.g., Carslaw and Jaeger, 1959, p. 136; Watanabe, 1992) for a transient heat transfer. If the sample assembly has a good cylindrical symmetry, a probe method also can be applied (Carslaw and Jaeger, 1959). As noted later in this paper, we tried both methods and obtained data consistent with other measurements only by the latter method. Instead of drilling a thin hole in the cylindrical core sample (10.0 mm in diameter and 20.0 mm in length), thin slits (0.50 mm) were cut into the cylinder. After placing the heater wire (0.1-mm nickel-chromium steel wire) and platinum-13 %rhodium thermocouple (0.1-mm wire), the slits were closed by gluing in thin plates cut from the original sample. The sample assembly, including the thermocouples and heater wire, was coated with silicone rubber to prevent penetration of pressure fluid during the experiment (Fig. 1). The whole sample assembly is compressed in a cylindrical vessel that creates hydrostatic pressures up to 1 GPa. The heater wire and thermocouple were connected to an electronic circuit controlled by a microcomputer (CPU) through a general purpose interface board (GPIB; Fig. 2 adapted from Kinoshita, 1992). Thermal diffusivity of the sample under pressure is obtained by applying a steady state heat-flow formula of a cylindrical symmetry (§7.2-V, Carslaw and Jaeger, 1959). A transient thermal transmission technique (formula 2.5.21, Beck et al , 1985) may be useful, but the latter method gave unrealistic values from our present experiment. The thermal conductivity is calculated on an assumption that the den-