Search

We have synthesized phase D at 24 GPa and at temperatures of 1250-1100 C in a multianvil press under conditions of high silica activity. The compressibility of this high-silica-activity phase D (Mg{sub 1.0}Si{sub 1.7}H{sub 3.0}O{sub 6}) has been measured up to 55.8 GPa at ambient temperature by powder X-ray diffraction. The volume (V) decreases smoothly with increasing pressure up to 40 GPa, consistent with the results reported in earlier studies. However, a kink is observed in the trend of V versus pressure above {approx}40 GPa, reflecting a change in the compression behavior. The data to 30 GPa fit well to a third-order Birch-Murnaghan equation of state (EoS), yielding V{sub o} = 85.1 {+-} 0.2 {angstrom}{sup 3}; K{sub o} = 167.9 {+-} 8.6 GPa; and K{prime}{sub o} = 4.3 {+-} 0.5, similar to results for Fe-Al-free phase D reported by Frost and Fei (1999). However, these parameters are larger than those reported for Fe-Al-bearing phase D and for Fe-Al-free phase D. The abnormal volume change in this study may be attributed to the reported hydrogen bond symmetrization in phase D. Fitting a third-order Birch-Murnaghan EoS to the data below 30 GPa yields a bulk modulus K{sub o} = 173 (2) GPamore » for the hydrogen-off-centered (HOC) phase and K{sub o} = 212 (15) GPa for the data above 40 GPa for the hydrogen-centered (HC) phase, assuming K{prime}{sub o} is 4. The calculated bulk modulus K{sub o} of the HC phase is 18% larger than the bulk modulus K{sub o} of the HOC phase.« less

[1] Brillouin scattering measurements have been carried out on an aqueous solution of Na2O-2SiO2 and anhydrous Na2O-2SiO2 glass and liquid under in situ pressures to 8 GPa in the temperature range from 300 to 850 K. This temperature range spans through the glass transition of both the hydrous and anhydrous forms of Na2O-2SiO2. The “modified” platelet scattering geometry has allowed us for the first time to determine VP independently from refractive index and hence the adiabatic compressibility and density of liquids as a function of pressure and temperature. The observed increase in density of the melt and glass phases formed at high-pressure-temperature conditions is likely associated with structural effects. There is a marked change in pressure derivative of adiabatic bulk modulus (K′S = dKS/dP) between the two phases. The large values of K′S of the liquid phase illustrate that the means of compaction of the liquid differ substantially from those of the glass and that the liquid is able to access a wider range of compaction mechanisms. The measured bulk modulus of the aqueous sodium silicate solution is highly temperature-dependent but is much more similar to values of silicate melts than to that of end-member water, particularly at high pressures. Thus water-rich silica-bearing solutions present at depth are likely to be difficult to seismically distinguish from anhydrous silicate melts solely on the basis of their sound velocities. The data on water-bearing compounds allow the pressure dependence of the partial molar volume of water to be assessed to upper mantle depths. These results are in excellent agreement with prior determinations and illustrate the utility of Brillouin experiments coupled with the externally heated diamond anvil cell in determining partial molar properties of liquids at high pressures.

Compressional and shear wave velocities (Vp and Vs) have been measured to 10 kbar in 17 granulite facies rocks and 15 eclogites. The former included quartzo-feldspathic, gabbroic, and garnet granulites as well as mangerites from the Adirondack region. The eclogites included the three types described by Coleman et al. from California, Norway, South Africa, and Tasmania. The mean ranges of the values of (∂Vp/∂P)T and (∂Vs/∂P)T in the 8- to 10-kbar range are slightly higher (0.015–0.022 and 0.008–0.012 km/s kbar) for eclogites than for granulites. Velocity-density systematics, based on the 10-kbar data, is evaluated in the light of Birch's law, D. L. Anderson's seismic equation of state, Wang's C-ρ relation, and the K-V relationship of O. L. Anderson and Nafe and D. L. Anderson and O. L. Anderson. On close analysis, there is a distinction in Wang's relations for M¯ ∼ 21 and M¯ ∼ 22. Most of the granulites and eclogites have M¯ values of ∼22; Birch's relationship for these and previously studied basalts ( M¯ ∼ 22) is Vp = −1.85 + 2.87ρ, where r2 = 96%. For Vs the best fit of the data for granulites and eclogites ( M¯ ∼ 22) is Vs = −0.33 + 1.40ρ, where r2 = 88%. The best-fit equation for Vs that includes the calcium oxide effect in all the granulites and eclogites is Vs = −0.63 + 0.21(21 − M¯) + 1.56ρ + 0.016 [CaO]. On the basis of the laboratory results, it is shown that elastic properties of garnet granulite with M¯ ∼ 22 are compatible with those of the 7.1- to 7.8-km/s crustal layer. Analysis of crust-mantle seismic data in various regions, and the present laboratory results, shows that the assumption of M¯ ∼ 22 for the lower crust as well as for the upper mantle fits better than M¯ ∼ 21. On the other hand, the M¯ ∼ 21 model fits better for some geophysically anomalous regions in the western United States.

Twenty-five gravity stations, based on Madison, Wisconsin (980.3689 gals), embracing a change of 1016 mgal along an approximately north-south line were established as a preliminary step in setting up a gravity standardization range in India. The instrument, a small LaCoste and Romberg geodetic gravimeter, model G-1-A, was transported mostly by air over a network between Trivandrum in the south and Srinagar in the north. Ten of the sites occupied are reoccupations of stations established earlier by the University of Wisconsin and the Survey of India, A comparison of the present results with the earlier measurements indicates either a tare or an elevation effect in one of the earlier sets of measurements which affects the previous connection to the national gravity base at Dehra Dun. The present connection to Dehra Dun gives a value of 979.0640 gals, which is 0.4 mgal higher than the earlier University of Wisconsin value and 0.8 mgal higher than the value used by the Survey of India.

The elastic constants of single-crystal rutile measured to 7.5 kb at 25°C and to 100°C at 1 bar, by the ultrasonic interferometry technique, are found to vary linearly with pressure and temperature within experimental error. The pressure and temperature derivatives of the elastic constants are:Cij dCij/dP dCij/dT, kb/degC11 6.47 −0.51C33 8.34 −0.90C44 1.10 −0.22C66 6.43 −0.78C12 9.10 −0.58C13 5.02 −0.33The value of (dKs/dP), where Ks is the isotropic adiabatic bulk modulus derived from the Voigt-Reuss-Hill (VRH) approximation, is 6.76; this value is higher than the values reported for other oxides of geophysical importance, such as MgO, Al2O3, and CaO but is close to the value for SiO2 (α quartz). The value of (dμH/dp), where μH is the isotropic shear modulus derived from the VRH approximation, is low (0.78); its geophysical implication is discussed. The values for (dKs/dT) and (dμH/dT) are −0.41 and −0.27 kb/deg, respectively. Significantly large differences between (dμv/dP) and (dμR/dP) and between (dμv/dT) and (Dμv/dT) are found in rutile because of its elastic anisotropy. The pressure and temperature dependence of the isotropic velocities are: (dνp/dP) = 7.64 × 10−3 km/sec/kb; (dνs/dP) = 5.20 × 10−4 km/sec/kb; (dνp/dT = −8.9 × 10−4 km/sec/deg; and (dνs/dT) = −3.9 × 104 km/sec/deg. The values obtained for the critical thermal gradients (∂T/∂Z)νs and (∂T/∂Z)ν s are, respectively, 4.64 and 1.74°C/km; the significance of the low (∂T/∂Z)ν s value is that no unreasonably high temperature gradient would be necessary to effect a decrease in shear velocity with depth. The reported shock-wave and X-ray data for rutile are discussed in the light of the ultrasonic equation of state.

The pressure-volume relationship for TiO2 (rutile) was determined under hydrostatic pressure up to 106 kbar at room temperature by means of X ray diffraction, employing the diamond anvil pressure cell and ruby fluorescence pressure calibration technique. Assuming an ultrasonic value of K0′ (6.84) from the Birch-Murnaghan equation, the zero-pressure bulk modulus (K0) value derived from the present pressure volume data is 2.11±0.07 Mbar, in good agreement with the ultrasonic isothermal Reuss bulk modulus values (2.076–2.095 Mbar).

Viscoelastic properties of basalt melts of a wide range of composition are determined from ultrasonic compressional wave velocity Vp and quality factor Qp measurements in the frequency range of 3.4–22 MHz and the temperature range of 1150°–1500°C. The basalt samples are grouped into two types: low magnesia (group A) containing 5–9 wt % MgO and high magnesia (group B) containing 18–23 wt % MgO. The Vp values of the melts determined at 1400°C are in the range 2.53–2.64 km/s for group A basalts and 2.79–2.86 km/s for group B basalts. The relaxed (low frequency) velocity Vr is a principally linear function of the melt density ρ. Least squares fits of the data give velocity-density relations Vr (km/s) = 0.691 + 0.712ρ (g/cm3) for group A basalts and Vr = 0.328 + 0.907ρ for group B basalts at 1400°C. The mean molecular weights M are 65.0 ± 0.3 and 58.2 ± 0.5 g/mol for groups A and B basalts, respectively, suggesting that M is also a principal factor determining the melt velocity. The relaxed modulus Mr of the basalt melts range from 16.5 to 22.8 GPa and is also a linear function of the melt density. The values of longitudinal ηL and volume ηv viscosities are calculated from the measured Vp and Qp and the available shear viscosity ηs data. The viscosities and related activation energies decrease almost linearly as NBO/T (nonbridging oxygens per tetrahedrally coordinated cation) increases.

The single-crystal elastic constants Cij of β-cubic PbF2 (fluorite type) have been measured to ∼0.3 GPa at ambient temperature and to 573 K at ambient pressure. The measured Cij are (in gigapascals (GPa)): C11 = 96.075 ± 0.094, CS = (C11 – C12)/2 = 24.475 ± 0.018, and C44 = 20.835 ± 0.012. The results from the least squares linear fit of the pressure data to ∼0.3 GPa yield dC11/dP = 7.79 ± 0.08, dCs/dP = 0.21 ± 0.08, and dC44/dP = 0.82 ± 0.02, and the temperature data in the range of 298–373 K give (in GPa per degree × 102): dC11/dT = −5.63 ± 0.05, dCS/dP = −1.26 ± 0.03, and dC44/dT = −1.11 ± 0.07. The pressure derivatives do not give any indication of the monotropic transition at ≤0.4 GPa; likewise, the temperature derivatives do not indicate the onset of superionicity at high temperatures. Relative stability of the α, β, and β′ phases of PbF2 is discussed in light of thermodynamic parameters.

Laboratory studies have been carried out to determine the causes of velocity and attenuation anisotropy in pelagic carbonate sediments from Deep Sea Drilling Project site 288 on the Ontong-Java Plateau. Compressional velocity Vp was measured under independently controlled pore pressure and confining pressure. Velocity anisotropy for compressional waves (horizontal velocity > vertical velocity) is found to decrease with increasing effective pressure, i.e., with closure of pores. Previous studies have suggested that velocity anisotropy in pelagic carbonate sequences is due largely to preferred vertical orientation of calcite c axes. The present results indicate that preferred horizontal orientation of grain contacts and flat pores is also a significant cause of the observed anisotropy for many samples. Significant anisotropy in ultrasonic attenuation for compressional (Qp−1) and shear (Qs−1) waves is observed such that the losses are greatest for propagation modes that can generate relative motion across horizontal grain contacts or flat horizontal pores. The results are discussed in terms of the possible mechanisms of dissipation.

An interferometric technique has been developed for the purpose of measuring ultrasonic velocity and attenuation in rock melts. In essence, a long wave train is transmitted through two long buffer rods separated by a thin layer of melt. Multiple reflections in the melt layer interfere with each other and give rise to resonances for melt thicknesses equal to integral multiples of half the wavelength. The velocity and attenuation can be obtained by measuring the amplitude of the transmitted wave as a function of either the melt layer thickness or the frequency. The accuracy of the method was tested at room temperature by compressional wave measurements on water, mercury, and carbon disulfide. Velocities were found to be accurate to within 0.3%. Values of the specific dissipation factor Q/sup -1/ as low as 10/sup -3/ can be resolved for a liquid such as water with a low acoustical impedance, whereas Q/sup -1/ values as low as 10/sup -4/ can be measured for high-impedance liquids such as mercury. Q/sup -1/ values obtained on carbon disulfide agree with previous work to within better than 0.001. Experiments on a viscous organic polymer show that shear measurements can be made and that Q/sup -1/ values upmore » to at least 0.2 can be measured by this technique. The method has been successfully used to study a basalt melt to 1500/sup 0/C. Sources of error and possible improvements are discussed, and the method is compared to previously used techniques. The method has the potential of measuring velocity and attenuation in partial melts and solids.« less

Electrical impedance in a partially molten gabbro was measured with the two-electrode method in the frequency range of 0.032 Hz to 10 kHz. Samples of different thicknesses were used to separate the bulk polarization effect from the sample-electrode interface polarization effect. The measurements show that both real and imaginary impedances increase linearly with increasing sample thickness. When the impedance data are extrapolated to zero thickness, the real impedance is as small as the experimental error, whereas the imaginary impedance has a finite value. This result indicates that the real impedance is mainly controlled by the electrical properties of the partially molten gabbro itself and that only small amounts of electrode interface polarization contribute to the imaginary impedance. After correction for the interface polarization effect, the present data were analyzed by the theoretical expression that describes the electrical impedance of partially molten rocks. The results show that frequency dispersion of electrical impedance in partially molten gabbro is associated with the polarization effect of ions in the partial melt.