Volume 84, Issue 3

This paper presents high-resolution synchrotron X-ray powder diffraction data at 290 K on two Fe-bearing. polycrystalline silicate perovskite samples with approximate compo­sitions (Mg 0.95 Fe 0.05 )SiO 3 synthesized at 25 GPa and 1920 K in a multi-anvil press at dif­ferent oxygen fugacity conditions. Mössbauer smdies have indicated that Fe 3 /∑Fe for the samples are 0.09 ± 0.01 and near 0.16 ± 0.03. Rietveld structural refinements confirm that Fe 2+ and Fe 3+ dominantly substitute for Mg 2+ in the 8-fold to 12-fold coordinated A site for both compositions. There appears to be no significant differences in the bond distances for these amounts of Fe 3+ and no conclusive structural evidence to support in­dications from Mossbauer experiments that Fe 3+ may occupy both A and B sites. To explore the effect of valence state further, this study also reports the first diffraction patterns of (Mg-Fe)SiO 3 perovskite collected at a wavelength near the Fe absorption edge.

This study presents evidence that for a solid solution of MgxFe x-1 SiO 3 perovskite, the shift in volume thermal expansivity, α, is small as the index χ is changed. According to data obtained theoretically by Hama and Suito (1998), α decreases by 0.3-0.4% as χ changes from 1 to 0.9 in a temperature range of 1900 K. Furthermore, the relative shift in α for Fe substitution is -0.4% under lower mantle conditions. Hama and Suito used the Vinet equation of state to calculate thermoelastic properties. Using a thermodynamic approach applied to the properties of a Debye solid, the data show that the relative shift in a is less than the relative change in specific volume, and, as χ changes from 1 to 0.9, V increases by 0.6% and a decreases by 0.4%.

We studied MgSiO 3 with the perovskite structure heated to temperatures up to 1500 K at pressures between 36 and 110 GPa with in-situ X-ray diffraction. The new pressurevolume-temperature (P-V-T) data were combined with literature data to provide thermal expansivity α and compressibility β against T (in K): α T =2.71 × 1O -5 + 1.80 × 1O -9 T - 1.48 T 2 (Model 1) or α T =2.13 × 1O -5 + 7.57 × 10 -9 T- 1.02 T 2 (Model 2), and β T = 3.735 × 10 -7 + 3.27 × 1O -11 T + 6.60 × 1O -15 T 2 . Model 1 yields physical properties of perovskite that confirm Anderson’s (1998) Debye approach; the model is valid for extrapolation to 3000 K or more. The parameters at 300 K are: α = 1.1 × 1O -5 , K 0 ′ (bulk modulus) = 261 GPa, K′ 0 = 4 and (∂K/∂T) p = -0.027. Thermal expansivity from this model does not fit the data of Funamori et al. (1996) at high temperature for P = 25 GPa. Model 2 uses an equation for α based on the data of Funamori et al. (1996), fits the available experimental data closely, and maintains conformity with Anderson’s Debye ap­proach. Heat capacity, C p , data for perovskite is given by either: C p = 110.8 + 8.031 × 1O -3 T - 1.302 × 1O -7 T 2 - 1.647 × 1O 7 T 2 + 2.755 × 1O 9 T -3 + 267.5 T -0.5 + 9287 T -1 (Model 1) or C p = 121.33 + 2.77 × 10 -3 T - 2.585 × 1O -6 T 2 - 1.710 × 1O 7 T + 2.792 × 1O 9 T -3 - 169 T -0.5 + 15782 T -1 (Model 2).

The phase diagram for MgSiO 3 was estimated using the entropy, enthalpy, thermal expansivity, and volumes of all the phases. Entropy at the various P-T conditions in the phase diagram was estimated using statistical thermodynamics and spectroscopic data at ambient and high pressures for each of the phases. Nearly complete 1 atm polarized Raman spectra of end-member MgSiO 3 orthoenstatite and new high-pressure Raman data on orthoenstatite (Pbca) to 24.5 GPa and majorite to 33.6 GPa are presented. Both of these minerals exhibit profound changes in their spectra as pressure is increased and the pressure dependence of the Raman modes changes substantially at 5 GPa for orthoenstatite and 26 GPa for majorite. These, like MgSiO 3 perovskite, appear to change symmetry even at room temperature. The slopes for the following transitions are reported: clinoenstatite (C2/c) to majorite, 212 bar/K; majorite to ilmenite, 46 bar/K; ilmenite to perovskite, 246 bar/K, majorite to perovskite, 26 bar/K. A volume change of 0.6 cm3/mol for the orthopyroxene to highpressure clinopyroxene transition was estimated using the previously measured phase boundary and the present entropy data. Clapeyron slopes are overestimated by 20 to 100% if the pressure dependence of ΔS across the transitions at various P-T conditions is not included in the thermodynamic calculations.

The crystal structure of (Mg 1.54 Li 0.23 Sc 0.23 )Si 2 O 6 protopyroxene has been studied with single- crystal X-ray diffraction at pressures to 9.98 GPa and Raman spectroscopy to 10.4 GPa. A first-order displacive phase transformation from the Pbcn space group to P2 1 cn was observed between 2.03 and 2.50 GPa, which is characterized by a discontinuous decrease in a, c, and V by 1.1, 2.4, and 2.6%, respectively, and an increase in b by 0.9%, along with appearance of intensities of some 0kl reflections with k ≠ 2n. This is the first substantiated example of protopyroxene having the symmetry predicted by Thompson (1970). Evidence for the phase transition from Raman spectroscopy is also presented. The prominent structural changes associated with the Pbcn-to-P2 1 cn transformation involve the abrupt splitting of one type of O-rotated silicate chain in low-pressure protopyroxene into S-rotated A and O-rotated B chains in high-pressure protopyroxene, coupled with a marked decrease in the O3-O3-O3 angles and a re-configuration of O atoms around the M2 site. The kinking angle of the silicate chain in the low-pressure phase at 2.03 GPa is 165.9°, whereas the angles are 147.9° and 153.9° for the A and B chains, respectively, in highpressure phase at 2.50 GPa. Strikingly, the two types of silicate chains in the P2 1 cn structure alternate along the b axis in a tetrahedral layer parallel to (100). Such a mixed arrangement of two differently rotated silicate chains in a tetrahedral layer has not been observed in any other pyroxene structure. Compression anisotropy of the protopyroxene structure is affected by the phase transition. The relative linear compressibilities (βa:βb:βc) are 1.00:1.72:0.99 for low-pressure protopyroxene, but are 1.00:1.28:1.65 for high-pressure protopyroxene. The bulk moduli of low- and high-pressure phases are 130(3) and 111(1) GPa, respectively. This study concludes that the Pbcn-to-P2 1 cn phase transition results from the differential compression between SiO 4 tetrahedra and MO 6 octahedra.

A new phase isostnictural with the minerals of the aenigmatite group Na 2 Mg 4+x Fe 3+ 2-2x Si 6+x O 20 with x = 0.25 - 0.5 was synthesized at 13-14 GPa with a split- sphere anvil apparatus (USSA-2000). The structure (for x = 0.4)_was determined from twinned-crystal X-ray diffraction data. The unit cell is triclinic. P1̅, a = 10.328(1), b = 10.724(1), c = 8.805(1) Å, α = 105.15(1), β = 96.85(1), γ =125.47(1)°, V = 719.67(3) A 3 , Z = 2. calculated density = 3.335 g/cm 3 . The twin law, independently determined from electron diffraction and transmission electron microscopy and by inspection of the X-ray data collected with an area detector, relates the twin components by a 180° rotation about [HO]*. Due to the coupled substitution. 2Fe 3+ = MgSi, which introduces octahedral Si. the stability of the phases with the aenigmatite structure apparently expands with in­creasing pressure. Hence, these phases could play a major role in the transition zone (410- 660 km), where the more common minerals they are replacing, olivine and clinopyroxene, reach the limit of then stability. The new evidence for the stability of aenigmatite-like minerals in the deep mantle could have important implications for the origin of the parental magmas producing aenigmatite-bearing and other agpaitic rocks.

Akimotoite, (Mg,Fe)SiO 3 , a new silicate mineral of the ilmenite group, was found in the shock-metamorphosed Tenham chondrite. It occurs as aggregates adjacent to clinoenstatite in fragments within shock-induced melt veins. Chemical analyses show the simplified formula to be (Mg 0.79 Fe 0.21 )SiO 3 ,the same as for clinoenstatite. Selected-area electron diffraction (SAED) patterns correspond to the synthetic (Mg,Fe)SiO 3 ilmenite phase with space group R3̅ . Lattice parameters derived from SAED patterns are a = 0.478(5) nm, c = 1.36(1) nm and V = 0.269(8) nm 3 in the hexagonal setting. The calculated density is 4.0(1) g/cm 3 . Akimotoite in this occurrence is thought to have been transformed from original orthoenstatite in a solid-state reaction produced by a shock event. Peak pressure and temperature generated by shock events in Tenham are estimated to be 22 GPa <Pmax <26 GPa and Tmax >2000 °C, assuming that equilibrium crystallization of aluminous majorite occurred in the melt veins. This new mineral was named after Syun-iti Akimoto.

Isothermal static compression data for MgO at 300 and 1100 K were obtained by combining synchrotron X-ray diffraction techniques with an externally heated high-temperature diamond-anvil cell that is capable of achieving pressures greater than 125 GPa at temperatures up to 1100 K. The experiments at 300 K were conducted under both hydrostatic and nonhydrostatic conditions. The deviatoric stress in non-hydrostatic environment significantly effects the measured lattice parameters. Fits to the static compression data of MgO at 300 K yield a bulk modulus of 185(7) GPa and 160(2) GPa under nonhydrostatic and hydrostatic conditions, respectively. The deviatoric stress decreases with increasing temperature, and a nearly hydrostatic condition was achieved at temperatures above 900 K when NaCl was used as a pressure-medium. The bulk modulus of MgO was determined to be 135(3) GPa at a temperature of 1100 K, yielding its temperature derivative of 20.030(3) GPa/K. Comparing these results with previous studies in the system MgO-FeO, shows that the bulk modulus of (Mg,Fe)O decreases with increasing FeO content, from 160 GPa for MgO to 146 GPa for FeO.

High-pressure single-crystal X-ray diffraction measurements of CaTiO 3 and CaGeO 3 perovskite have been carried out to 9.7 and 8.6 GPa, respectively, at room temperature. Fitting a third-order Birch-Murnaghan equation-of-state to the P-V data yields values of V 0 = 223.764 ± 0.017 Å 3 , K T,0 = 170.9 ± 1.4 GPa, and K′ = ∂K/∂P = 6.6 ± 0.3 for CaTiO 3 and V 0 = 206.490 ± 0.017 Å 3 , K T,0 = 194.0 ± 2.1 GPa, and K′ = 6.1 ± 0.5 for CaGeO 3 . A similar analysis of the axial compressibilities shows that the degree of anisotropic compres­sion in both perovskites is less than 10%. In CaTiO 3 the a and b axes have similar com­pressibilities (K a = 168.7 ±2.1 GPa, K b = 168.3 ±1.9 GPa) whereas the c axis is the least compressible (K c = 175.3 ±1.5 GPa). In CaGeO 3 , the b axis (K b = 188 ± 4 GPa) and the a axis (K a = 195 ± 5 GPa) are more compressible than the c axis (K c = 204 ± 3 GPa). The variations with pressure of all axes show significant curvature with increasing pressure and have K′ values ranging from 5.7 ± 0.5 to 7.0 ± 0.4 in CaTiO 3 and 5.0 ± 0.9 to 6.9 ± 1.2 in CaGeO 3 . No phase transition was detected. There is evidence, however, that in CaGeO 3 the tetragonal to orthorhombic spontaneous strain decreases slightly with pressure which may indicate that a phase transition occurs at a pressure above 10 GPa. Elasticity trends of Ca-perovskites relating bulk modulus and molar volume are independent of both the degree of distortion from cubic symmetry and the symmetry of the structure.

The equations of state of A2/a titanite phases of CaTiSiO 5 , Ca(Ti 0.5 Si 0.5 )SiO 5 , and CaSi 2 O 5 have been determined from high-pressure X-ray diffraction measurements. The isothermal bulk moduli are K 0,T = 131.4(7) GPa (for P >3.6 GPa), 151.9(1.6) GPa, and 178.2(7) GPa, respectively, for a second order Birch-Murnaghan equation-of-state (i.e., with K′=∂K∂/P = 4). Refinements of third order equations-of-state yielded values of K′ that did not differ significantly from 4. The complete substitution of Si for Ti in the octahedral sites of the titanite structure, therefore, results in approximately a 30% increase in bulk modulus and a 13% increase in density. The large stiffening of the structure can be attributed to the absence of rigid-unit modes from the structure and the direct involvement of cationcation interactions in the compression of the structure.

We present a first-principles computational study of the structural properties of ringwoodite and the influence of cation exchange on these properties as well as the enthalpy differences between ringwoodite and inverse ringwoodite. Our results agree with lowtemperature experiments, in that cubic ringwoodite is the stable structure up to 25 GPa. This pressure range encompasses the lower part of the mantle transition zone where ringwoodite is thought to be the most abundant phase. The equation of state as derived from experiment and theory are in good agreement. In contrast to normal ringwoodite, inverse ringwoodite compresses highly anisotropically and the compression mechanisms differ considerably for the two structures. The predicted enthalpy difference between normal and inverse ringwoodite is significantly smaller than that obtained from a simple ionic model (O’Neill and Navrotsky 1983), emphasizing the importance of including structural relaxation in models that address the energetics of order-disorder reactions, at least in the case of ringwoodite.

Synchrotron anomalous scattering experiments with X-ray wavelengths at the Fe K absorption edge have made it possible to distinguish between ions having different valence states in magnetite at low temperatures. Using a large difference in anomalous scattering factors between Fe 2+ and Fe 3+ , the valence-difference contrast method was applied for the intensity measurements of fundamental and superlattice reflections of magnetite at low temperature. Reflections with half-integer indices, such as 0 4 7⁄2 and 4 4 9⁄2, were observed at 102 K, showing that magnetite transformed to a lower symmetry form. The energy dependence in the diffracted intensity was clearly observed at the Fe K-edge for the 0 4 7⁄2 superlattice reflection at 102 K. A sharp minimum of the normalized intensity at E = 7.122 keV was explained by the valence contrast between Fe ions occupied in the respective upper and lower parts of the double cell. Thus, our results strongly suggest the existence of charge ordering between ferrous and ferric ions in magnetite at low temperatures.

The temperature dependence of the cation distribution in synthetic spinel (MgAl 2 O 4 ) was determined using in-situ time-of-flight neutron powder diffraction. Neutron diffraction patterns of stoichiometric MgAl 2 O 4 and slightly non-stoichiometric Mg 0.99 Al 2 O 4 samples were collected under vacuum on heating from room temperature to 1600 °C, and the cation distribution was determined directly from site occupancies obtained by Rietveld refinement. The equilibrium non-convergent ordering has been analyzed using both the O’Neill-Navrotsky and Landau thermodynamic models, both of which fit the observed behavior well over the temperature range of the measurements. Fitting the data between 560 °C and 1600 °C using the O′Neill and Navrotsky (1983) thermodynamic model yields α = 32.8 ± 0.9 kJ/mol and β = 4.7 ± 2.0 kJ/mol. The fit to the Landau expression for ordering gives values of T c = 445 ± 109 K and c′ = 1.62 ± 0.21. This confirms suggestions that the sign of the β coefficient in FeAl 2 O 4 and MgAl 2 O 4 is positive, and opposite to that found in other 2-3 oxide spinels. Non-equilibrium order-disorder behavior below 600 °C has been analyzed using the Ginzburg-Landau model, and successfully explains the time-temperature dependent relaxation behavior observed in the inversion parameter. Changing the stoichiometry, even by as little as 1 mol% Mg-deficiency, significantly reduces the degree of order.

The cluster variation method (CVM) for determining the thermodynamics of both shortrange and long-range order in this study was extended to complex aluminosilicate minerals. Using recently developed algorithms for the maximization of cluster entropies, it is possible to construct high-order cluster variation approximations for Ising lattices of complex topology. This study gives the essentials of developing CVM models for the Al, Si nets of nepheline, leucite, feldspar, and cordierite. The robustness of the method is demonstrated by its ability to reproduce quantitatively recently published results of Monte Carlo simulations of Al, Si ordering in the Ising model with feldspar topology. The future prospects for the application of CVM to aluminosilicates are discussed.

Structure deformations induced by pressure and temperature in synthetic “Cs-tetra-ferri-annite” 1M [Cs 1.78 (Fe 2+ 5.93 Fe 3+ 0.07 )(Si 6.15 Fe 3+ 1.80 Al 0.05 )O 20 (OH) 4 ], space group C2/m, were an­alyzed to investigate the capability of the mica structure to store the radiogenic isotopes 135 Cs and 137 Cs. “Cs-tetra-feni-annite” is not a mineral name, but for the sake of brevity is used here to designate a synthetic analog of the mineral tetra-ferri-annite. The bulk modulus and its pressure derivative determined by fitting the unit-cell volumes between 0 and 47 kbar to a third-order Birch-Mumaghan equation of state are K 0 = 257(8) kbar and K 0 ′ = 21(1). respectively. Between 23 °C and 582 °C. the a and b lattice parameters remain essentially unchanged, but the thermal expansion coefficient of the c axis is a c = 3.12(9) x 10 -5 ° C -1 . High pressure (P) and high temperature (T) produce limited internal strain in the structure. The tetrahedral rotation angle, α , is very small and does not change significantly throughout the P and T range investigated. Above 450 °C in air. “Cs-tetra- ferri-annite" underwent an oxidation of octahedral iron in the M2cis site, balanced by the loss of H and shown by a decrease of the unit-cell volume. Independent isobaric data on thermal expansion and isothermal compressibility data define the “geometric” equation of state for “Cs-tetra-ferri-annite”: V/V 0 = 1 + 3.0(1) 1O -5 T - 2.68(9) 1O -3 P + 2.0(2) × 1O -5 P 2 where T is in degrees Celsius, P is in kilobars. The α / β ratio of about 12 bar/°C indicate that the cell volume of “Cs-tetra-ferri-annite” remains unchanged under geothermal gradients of ~23 °C/km. On the whole, the data confirm that the structure of “Cs-tetra-ferri-annite” may be a suitable candidate for the storage of large ions, such as Cs in the interlayer and should be considered as a potential Synroc component.

Structural and volume compressibility data for reedmergnerite, NaBSi 3 O 8 , were obtained by single-crystal X-ray diffraction at pressures up to 4.7 GPa. The bulk modulus was determined to be 69.8(5) GPa with the pressure derivative constrained to 4. Unit-cell compression is anisotropic, as indicated by unit strain tensors. Tetrahedral bond lengths and angles remained relatively constant over the pressure interval, whereas Na-O bonds decreased systematically. T-O-T angles underwent a variety of behaviors, remaining constant or decreasing with pressure. The compression for reedmergnerite is similar to that of low-albite, wherein bending of the (Al,B)-Oco-Si angle compresses the Na-bearing zigzag channels. In contrast, microcline compresses by shearing the four-membered rings, which in turn compresses the Kbearing channels. At about 4 GPa, a new bond between K and Obm appears that alters the compression mechanism and explains the discontinuity in the pressure variation of crystallographic parameters observed by Allan and Angel (1997). Thus, the compression mechanism of the alkali feldspars is dominated by the compression of alkali containing channels. However, because of low symmetry, this can be accomplished in several ways. The observed variety of compression pathways resulted from T-O-T angle bending energetics that were coupled with the effects of alkali cation bonding.

Molten NaCl was analyzed by high-pressure X-ray diffraction experiments using synchrotron radiation up to 5 GPa and 1600 °C along the melting curve. The interference function Qi(Q) and the correlation function g(r) were derived from the diffraction data. The first-neighbor distance r 1 is about 2.7 Å and the second-neighbor distance appears around 1.4r 1 -1.5r 1 . The coordination number, CN, of the nearest neighbor ions increases with pressure from 3.5 at 0.1 MPa to 4.5 at 5 GPa. This is the evidence that the NaCl melt has a B1-like structure with large vacancies over this pressure range and becomes densified by an increase in CN as a result of second neighbor compaction.

Molecular Dynamics (MD) simulations were used to study the relationship between structure and transport properties in five liquids (NaAlO 2, Na 2 Al 2 SiO 6, NaAlSiO 4, Na-AlSi 2 O 6, and NaAlSi 3 O 8 ) in the system NaAlO 2 -NaAlSi 3 O 8 at temperatures ranging from 4000 to 6000 K and pressures from 0 to 55 GPa. Seventy simulations were earned out in the microcanonical ensemble using a simple pair-wise additive potential with Coulombic interaction and Born-Mayer repulsion. Detailed study of the coordination of O and network forming cations provides a master set of coordination environment or speciation curves. These master curves were applicable to all compositions and temperatures and were most explicit when compression (V r/ V; where V r is the molar volume at a reference pressure) was used as the independent variable. The universality implied that coordination environ­ments for network atoms 0. Al. and Si depend weakly upon Si/Al, T/O, or Na/T atomic ratios for the compositions studied. Self-diffusion coefficients, computed from analysis of mean-square displacements, were used to evaluate the activation enthalpy (H a = E a + PV a ) for self-diffusion for each species. The activation energy (E a ) for Na was independent of composition, whereas E a for O, Si, and Al increased as Si/Al increased. Activation volume (V a ) at pressure <15 GPa was positive for Na and negative for O, Si, and Al and decreased with increasing Si/Al for all species. An extension of the Adam-Gibbs-DiMarzio config­urational entropy theory taking explicit account of [2] O and [3] O mixing explained both the variation of the pressure-derivative of the shear viscosity as a function of composition and the disappearance of “anomalous” viscosity behavior at P > ~25 GPa for all compositions in the system NaAlO 2 -NaAlSi 3 O s .

Subsolidus and melting experiments of a K-rich basaltic composition have been performed to 27 GPa. Pyroxene and majoritic garnet contain significant amounts of K 2 O (1- 2 wt%) at 6-12.5 GPa and above 22.5 GPa, respectively. Both pyroxene and garnet may act as main reservoirs for K in the mantle. Sanidine (KAlSi 3 O 8 ) or its polymorph potassium hollandite is the first phase to be melted up to 20 GPa, but potassium hollandite becomes a liquidus phase at P ≤ 22.5 GPa. In contrast to its typical incompatible behavior at low pressures, potassium is a compatible element in the deep mantle, at least in basaltic systems. This compatible behavior implies that a K-rich melt like kimberlite is difficult to generate at lower mantle conditions.

Lawsonite is an important water reservoir in subducting oceanic crust below the amphibole dehydration depth ~70 km. To determine the maximum pressure stability of lawsonite in the MORB+H 2 O system, experiments were carried out using a 1000 ton uniaxial multi-anvil apparatus (SPI-1000). Mixtures of synthetic gel+2wt% H 2 O were used for the starting materials with the average MORB composition. Experimental P-T conditions were T = 700-900 8C and P = 5.5-13.5 GPa. Run durations were 12 and 24 h. Lawsonite was synthesized stably up to 10 GPa and T < 700 °C in the stishovite stability field, and <900 °C at 8 GPa and 750 °C at 5.5 GPa in the coesite stability field, with a steep positive slope for the lawsonite-out reaction. The lawsonite-out reaction in the coesite stability field changes to have a gentle negative slope in the stishovite stability field. The reaction leading to the disappearance of lawsonite is a continuous reaction due to the compositional enlargement of garnet toward the grossular end-member with increasing T and P. Lawsonite disappears when the tie line connecting grossular-rich garnet with omphacitic clinopyroxene reaches the bulk composition on the conventional AC(FM) ternary diagram.

High-pressure Brillouin scattering experiments were conducted on three natural, singlecrystal garnets with near end-member compositions: pyrope (Mg 3 Al 2 Si 3 O 12 ), grossular (Ca 3 Al 2 Si 3 O 12 ), and andradite (Ca 3 Fe 3+ 2 Si 3 O 12 ). Acoustic velocities of the samples were measured from the samples at pressures to 10 GPa in a diamond-anvil cell with a methanolethanol- water pressure medium. Elastic stiffness constants (C ij ) and their pressure dependencies (∂C ij /∂P) were calculated, as were aggregate elastic moduli (K S and G) and their respective pressure derivatives. The Cauchy relations changed with pressure for both grossular and andradite, suggesting that assuming these are constant is not valid for extrapolation to high pressures for these and possibly other minerals. The variation in ∂KS /∂P and ∂G/∂P was sensitive to composition and required a complex mixing model for extrapolation of end-member elasticity values from garnets of intermediate compositions.

Elastic wave velocities of Mg 3 Al 2 Si 3 O 12 pyrope garnet were measured to 10 GPa at ambient temperature, using ultrasonic interferometry in a 1000 ton split-cylinder, multi- anvil apparatus (USCA-1000). The sample used in the ultrasonic measurements was a polycrystalline specimen hot-pressed at 5 GPa and 1350 °C in a 2000 ton uniaxial split- sphere apparatus (USSA-2000) from a homogeneous glass of pyrope composition. Special P-T paths used during synthesis minimized effects of decompressing and thermal cracking; the bulk density of the sample was indistinguishable from the X-ray density. The elastic wave velocities measured at the ambient condition agree with the Hashin-Shtrikman av­erages of the single crystal values within the mutual uncertainties. The high-pressure ex­periments yielded the elastic moduli and their pressure derivatives (finite strain fit) for the shear modulus G 0 = 92 ± 1 GPa, G′ 0 = (∂G/∂P) T = 1.6 ± 0.2 and for the longitudinal modulus L 0 = 294 ± 1 GPa, L′ 0 = (∂L/∂P) T = 1A ± 0.5, (L = K s + 4/3G), from which the bulk modulus K 0 = 171 ±2 GPa, K′ 0 = (∂K/∂P) t = 5.3 ± 0.4 was calculated. The pressure derivative for the shear modulus of pyrope did not differ from those of natural pyrope-almandine-grossular garnets. For the bulk modulus, the pressure derivative for py­rope agreed with that for pyrope-almandine but was substantially higher (25%) than that for the Ca-bearing garnet. In the pyrope-majorite series, K′ 0 remained constant, whereas G′ 0 increased by 25 for 38% majorite content.

The silicon-free end member of the hydrogrossular solid solution series, katoite Ca 3 Al 2 [(OH) 4 ] 3 , was synthesized in a solid-media, piston-cylinder apparatus. The enthalpy of formation from the component oxides was measured by high-temperature oxide melt calorimetry and found to be ΔHf = 2255.6 ± 12.2 kJ/mol; the resulting enthalpy of formation from the elements is ΔHf = 25551.5 ± 16.4 kJ/mol. From this value, enthalpies for breakdown reactions of hydrogrossular were calculated and the relative energetic stability of katoite evaluated.

The systematics of Co and Ni in olivine from six Apollo 12 olivine basalts were studied by SIMS techniques and correlated with major and minor element data obtained with the electron microprobe. Our results, together with previous studies, demonstrate that one of these basalts (12009) was extruded, as a liquid, onto the lunar surface and was parental to five other cumulates (12075, 12020, 12018, 12040, and 12035). The concentrations of Ni in zoned olivine phenocrysts behave as expected for a highly compatible element with initial estimated D Ni = 9.9. On the other hand, Co concentrations in olivine vary hardly at all and show flat patterns across crystals that retain normal zoning trends for Mg, Fe, Mn, and Ni. The explanation for this behavior is not that D Co ≈ 1 (our estimated D Co = 4) or that Co zoning has been erased by rapid diffusion of Co compared to the other elements (e.g., Mg and Fe) that still show normal zoning. The explanation for the behavior, as originally suggested by Kohn et al. (1989), is that the increase in D Co with crystallization exactly balances the depletion of Co in the melt. This decoupling of the behavior of Ni and Co in olivine results in significant increases in Co/Ni with crystallization.

The crystal structure of gillulyite Tl 2 (As, Sb) 8 S 13 was reinterpreted based on the average structure determined by Foit et al. (1995). It consists of alternating PbS-like slabs comprised of As-S polyhedra and Tl-bearing slabs with partly zeolitic properties. In the latter, TlS 5 and As 2 S 5 groups alternate regularly along the b direction and across the width of the slab, thus eliminating the need for S-S and As-As bonds postulated by Foit et al. (1995). These two types of slabs are also unit order-disorder (OD) layers, the ambiguities in stacking of which lead to several potential gillulyite polytypes. The actually observed OD phenomena are caused by ambiguity in the position of [010] TlS 5 -As 2 S 5 sequences (or of entire such layers), which can with equal probability assume two positions 1b apart.

The uptake of Cu and Cd in aqueous solution by interaction with the surfaces of carefully synthesized finely sized particles of goethite, lepidocrocite, mackinawite, and pyrite was measured as a function of initial metal concentration in solution. The results show how reactions between metal ions in solution and mineral surfaces depend in a subtle way on the nature of the surfaces and, in certain cases, on the initial concentration of metal in solution. The uptake curves fall into two groups; type I in which the efficiency of uptake decreases with increasing concentration (Cu and Cd on goethite and lepidocrocite, Cd on pyrite), and type II in which it remains constant (Cu on mackinawite and pyrite, Cd on mackinawite). The total uptake is an order of magnitude greater for the latter group. X-ray absorption spectroscopies (XANES and EXAFS) were used to define the local environments of the metals taken up at the mineral surfaces. All examples showing the type I behavior yield information on local environments consistent with their being bound to the surfaces by an inner sphere complex formation mechanism. Thus, Cu on goethite appeared to form a Jahn-Teller distorted octahedral complex (four O atoms at 1.94 Å ; two O atoms at 2.41 Å ) but with evidence for interaction with two further Fe (or Cu) at 2.92 Å . On lepidocrocite, the first coordination sphere was essentially identical to that on goethite but with a second Cu or Fe shell at 3.04 Å and, for the highest Cu loadings, a third shell (2 Cu or Fe) at 3.67 Å . The Cd on goethite showed best fits for sixfold coordination to O (at 2.26 Å ) but with evidence for a second shell of Fe atoms at 3.75 Å . On lepidocrocite, the first shell was essentially the same as for goethite, but a second shell of Fe atoms appeared to occur at the shorter distance of 3.31 Å . For Cd-loaded pyrite, best fits were given by a single shell of six O atoms at 2.27 to 2.28 Å and with no evidence for a second shell of metal atoms. The systems exhibiting the type II behavior yielded spectra consistent with the formation of new phases on the surfaces, either by precipitation or replacement reactions. In the case of Cu interaction with mackinawite, a chalcopyrite phase appeared to form, whereas interaction with pyrite seemed to produce binary Cu sulfides-covellite at lower loadings and chalcocite at higher loadings. Cd interaction with mackinawite seemed to produce a CdS phase.

We investigated smelter-contaminated soils from Evin-Malmaison, Nord-Pas-de-Calais, France, and mine tailings from Leadville, Colorado, U.S.A. Bulk Pb concentrations range from 460 to 1900 ppm in the topsoils at Evin-Malmaison site and from 6000 to 10000 ppm in the tailings samples from the Leadville site. These concentrations necessarily raise human health and environmental concerns, but bioavailability and chemical lability of Pb in these materials vary dramatically and show little correlation with bulk concentrations. This study provides detailed information on the speciation of Pb in these materials. Emphasis is on the identification and characterization of poorly crystalline and/or fine-grained species, such as sorption complexes and poorly crystalline (co)precipitates, which are likely to control Pb bioavailability and mobility in these natural systems. Because these samples are heterogeneous, multi-phase mixtures, a variety of bulk analytical methods were used including powder X-ray diffraction (XRD) and Rietveld refinement, scanning electron microscopy, electron probe microanalysis, synchrotron-based Xray absorption, and micro-fluorescence spectroscopies. The synchrotron-based techniques enable identification of amorphous or nanocrystalline Pb-containing phases and the spatial distribution of Pb at the 25 μm scale. These techniques, in conjunction with physical and chemical separation techniques, allowed identification and characterization of several species not amenable to detection by conventional microanalytical techniques. In the Evin-Malmaison samples, direct spectroscopic evidence for Pb sorbed to humic acids was found, as well as to both manganese and iron (oxyhydr)oxides. In the Leadville samples, variations in Pb speciation with pH are consistent with predictions based on simplified model system studies of adsorption processes; specifically, the carbonate-buffered tailings with near-neutral pH contain up to 50% of total Pb as adsorption complexes on iron (oxyhydr)oxides, whereas Pb speciation in sulfide-rich low pH samples is dominated by Pb-bearing jarosites with no evidence for adsorbed Pb in these latter samples.

Topological and bond critical point properties of electron density distributions, ρ(r), were calculated for a series of sulfide molecules, containing first- and second-row main group M-cations. Laplacian maps of the distributions, ∇ 2 ρ(r), show that the valence shell charge concentration (VSCC) of the sulfide anion is highly polarized and extended into the internuclear region of the M-S bonds, coalescing with the VSCCs of the more electronegative first-row cations. On the other hand, maps for a corresponding set of oxide molecules show that the oxide anion tends to be less polarized and more locally concentrated in the vicinity of its valence shell, particularly when bonded to second-row M-cations. A search for extrema in the ∇ 2 ρ(r) distributions reveals maxima in the VSCCs that can be ascribed to bonded and nonbonded electron pairs. The different and distinctive properties of sulfides and oxides are examined in terms of the number and the positions of the electron pairs and the topographic features of the Laplacian maps. The evidence provided by the electron density distributions and its topological properties indicates that the bonded interactions in sulfides are more directional, for a given M-cation, than in oxides. The value of the electron density distribution at the bond critical point and the length of a given M-S bond are reliable measures of a bonded interaction, the greater the accumulation of the electron density and the shorter the bond, the greater its shared (covalent) interaction.

The 57 Fe nuclear forward scattering (NFS) of synchrotron radiation and the use of dia­mond anvils with helium as pressure medium allowed study of the electronic state of Fe 2+ in the chain silicate hedenbergite CaFeSi 2 O 6 at pressures up to 68 GPa. Characteristics of NFS time spectra were compared with those of conventional Mössbauer spectra. NFS time spectra of 57 Fe in hedenbergite revealed a reversible phase transition between 53 and 68 GPa at room temperature, which is probably a transition from the paramagnetic phase at low pressures to a magnetic phase at high pressures. If this interpretation is correct, the Néel temperature T N of hedenbergite depends critically on pressure (T N = 45 K at I atm).

Changes in configuration and hydrogen bonding are inferred from the complex responses of peak parameters (frequency, width, height, and area) to compression for various hydroxyl groups in dense, hydrous magnesium silicates (DHMS). Stretching frequencies (V i ) depend not only on both O-H...O bond length but also bond angle, which suggests a need to re-examine frequency- structure relationships at 1 atm. For phase B [Mg 12 vi Si iv Si 3 O 19 (OH) 2 ]. v of the most intense OH - peak determined from IR spectra at nearly hydrostatic pressures (P) decreases to a minimum at 5 GPa. and then rises to a broad maximum near 35 GPa, whereas v of the other intense peak de­creases to a broad minimum near 30 GPa. This behavior shows that compression of phase B ini­tially moves O19 toward H2 on O21 and rotates H1 about O20 away from 019 and toward 021 inside the triangle of O atoms, thus increasing the H1-020-019 bond angle (from 12 to 21° at 37 GPa). H-H repulsion rotates H2 outside of the triangle by 11° at 37 GPa. Changes in O-O bond length and H-O-O angle calculated from the trends of OH frequency with increasing pressure are consistent with the relatively incompressible layer of SiO 4 tetrahedra. Widths, areas, and heights of the hydroxyl peaks also increase with P (whereas areas of the lattice modes are constant), with H1 being much more affected: This response suggests that the H1 bond bends further with com­pression. Raman spectra of stoichiometric OH - modes in superhydrous phase B [Mg 10 vi Si rv Si 2 O 14 (OH) 4 ] roughly parallel these trends, consistent with similarly paired hydroxyls with bent bonds. For phase A [Mg 7 ivr Si 2 O s (OH) 6 ], the Raman OH - bands depend linearly on P. and with constant peak height with slopes similar to those observed for weak stretching modes of trace H in phase B. This behavior is common and is associated with linear hydrogen bonds and unpaired OH" groups. Combining our data with previous results reveals systematic, linear relationships of ∂v/∂P| i0 with v that suggest high-polyhedral bulk moduli for DHMS. Most frequencies of 31 mid-IR lattice modes for phase B depend linearly on curvature is seen for several modes. Raman frequencies of lattice modes from superhydrous phase B (shyB) and phase A depend linearly on P.