Volume 86, Issue 1-2

Laboratory experiments have been conducted to determine simultaneously the self diffusivities of Si and O in synthetic dacite melt (NBO/T = 0.1) from 1 to 5.7 GPa and from 1355 to 1662 °C. Glasses enriched in 18 O and 28 Si were synthesized and mated to their isotopically normal counterparts to form diffusion couples used in the piston cylinder device (1 and 2 GPa) and multi-anvil apparatus (4 to 5.7 GPa). Profiles of isotope abundances were measured by secondary ion mass spectrometry. Self-diffusion coefficients for Si (D* Si ) are significantly lower than self-diffusion coefficients for O (D* 0 ) at all run conditions; for example, D* 0 = 6.45 ± 0.65 × 10-14 m 2 /s and D* Si = 1.45 ± 0.45 × 10 -14 m 2 /s at 1 GPa and 1355 °C. The temperature dependence is similar, but not identical, for Si and O self diffusion at all pressures, yielding activation energies of 293-380 kJ/mol at 1 GPa, 264-305 kJ/mol at 2 GPa, and 155-163 kJ/mol at 4 GPa. The pressure dependence is similar for Si and O at all temperatures, giving activation volumes for Si and O that are -14.5 to -17.1 cm 3 /mol at 1460 °C, -9.8 to -8.7 cm 3 /mol at 1561 °C, and -8.8 to -9.3 cm 3 /mol at 1662 °C. Self-diffusion coefficients for Si and O reach maximum values at roughly 5 GPa. The mode of Si and O self diffusion in dacitic liquids is constrained by the large activation volumes, D* 0 ≈ 2 D* Si , and predictions using the Eyring equation, which suggest that Si and O diffuse as molecular species at 1460 °C. At 1561 and 1662 °C, less negative activation volumes and predictions of the Eyring equation are consistent with diffusion of Si and O by a combination of mechanisms, including the formation of a high-coordinated intermediate species.

The effect of anhydrous composition on the solubility of water in granitic melts was investigated experimentally at 800 °C and pressures from 50 to 500 MPa. Starting materials were ten natural obsidians from various localities worldwide and one re-melted leucogranite from the Himalayas. Most of the experiments were performed in externally heated pressure vessels using Ni-NiO to buffer f O₂ . All samples were quenched isobarically after reaction for 120-336 h. Water contents of the resulting glasses were determined by Karl-Fischer titration. The solubility data indicate that Na/K ratio and normative Qz content have only a minor effect on water solubility, whereas the (MCLNK-A)/O parameter, defined as 100·(2Mg + 2Ca + Li + Na + K- Al)/total oxygen, has a major effect. A parabolic law expressed as the mole fraction of H 2 O in the melt on a one-oxygen mole basis is proposed to describe the compositional dependence of water solubility in the range 50-200 MPa: X H₂O = X H₂O 0 · (1 + 0.05 ·{[(MCLNK-A)/O] - 0.5} 2 ) Minimum mole fractions of water in the melt (X H₂O 0 ) are 0.0521 at 50 MPa, 0.0757 at 100 MPa, and 0.1069 at 200 MPa. The equation fits water solubility data for granitic and phonolitic melts at 100 MPa and 200 MPa to within ±4% relative. The effects of anhydrous composition on water solubility are much more pronounced at 500 MPa than at lower pressures. Thus, the following expression was derived to represent the effects of anhydrous melt composition on water solubility at 500 MPa: X H₂O = 0.1681·(1 + 0.13·{[(MCLNK-A)/O] - 0.5} 2 ).

Zinc feldspar [Ca(ZnSi 3 )O 8 ] has been synthesized hydrothermally and its phase relations have been investigated in the system CaO-ZnO-SiO 2 up to pressures of 1.0 GPa. Single-phase Zn feldspar grew from an oxide mixture at 600 °C and 0.7 GPa within 6 days. It crystallizes in space group P - 1 displaying lattice constants of a = 8.121(1) Å, b = 12.927(1) Å, c =7.206(1) Å, α = 93.76 (5)°, β = 116.120(7)°, γ = 84.368(7)°, V = 675.7 Å 3 , Z = 2 (Heuer et al. 1998). In Zn feldspar, Zn and Si are ordered on the tetrahedral sites and the structure of Zn feldspar is more similar to the low albite structure than to that of anorthite due to the tetrahedral ordering scheme. Zinc feldspar does not melt congruently but decomposes at subsolidus conditions according to the reaction: 4 Zn feldspar = willemite + 2 hardystonite + 7 quartz (1). Reaction 1 was determined by means of reversed experiments and its negative P-T slope can be expressed by the equation P (GPa) = 2.797(0.325) - 0.0031 (0.0004) T (°C). Introducing the stability relations of the SiO 2 system, two invariant points, I qtz and I trid , are generated at 720 °C, 0.57 GPa and 896 °C, 0.016 GPa projected on the low/high-quartz and high-quartz/tridymite phase transitions, respectively. The phase assemblage willemite+ hardystonite+quartz is not stable at pressures above 1.2 GPa where petedunnite is formed according to the reaction: willemite + 2 hardystonite + 3 quartz = 4 petedunnite (2) (Rothkopf and Fehr 1998). The stability field of Zn feldspar is restricted to pressures <1.2 GPa and temperatures <530 °C, where Zn feldspar decomposes according to the doubly degenerated reaction: Zn feldspar = petedunnite + quartz (3). The intersection of reactions 1 to 3 generates an invariant point at ca. 530 °C and 1.2 GPa. Zinc usually is not determined routinely in electron microprobe analyses of feldspars and therefore data on the Zn contents of natural feldspar are scarce. Nevertheless, some feldspars of Pb-Zn hydrothermal vein deposits and skarns display deficiencies of tetrahedral cations and contain up to 0.18 wt% ZnO in combination with appreciable BaO contents, implying the existence of a Ba(ZnSi 3 )O 8 component in addition to Ca(ZnSi 3 )O 8 . We recommend inclusion of Zn in analysis of feldspars from such paragenesis.

Magmatic Na-rich phlogopite (1-5 wt% Na 2 O) is present as a late-crystallizing mineral in two groups of texturally and mineralogically distinct gabbroic xenoliths at Volcán San Pedro (36°S, Chile), an Andean arc volcano. Phlogopites are characterized by high 100·Mg/(Mg + Fe) (up to 83) and high Cr 2 O 3 contents (up to 0.4 wt%), and they are always found surrounding variably resorbed olivine, pyroxenes, Cr-spinel, and in some cases, plagioclase. We interpret these micas as the result of open-system processes involving infiltration of water-rich evolved melts [with high Na/(Na + K)] and reaction with refractory minerals. The highest 100·Na/(Na + K) (~70) and Na 2 O concentrations (~5 wt%) in phlogopite appear to require reaction with liquids of unrealistically high Na/(Na + K) if no other factor is considered. This, together with the observation that phlogopites consist of alternating Na-rich and Na-poor cleavage-parallel bands, can be best interpreted by the presence of a solvus between the aspidolite (Na) and phlogopite (K) end-members. The high proportions (up to 15 vol%) of Na-rich phlogopite in two different groups of gabbroic xenoliths suggest that it might be a more common and abundant mineral than has been previously recognized, and that it may be used as an indicator of open-system processes.

Two samples from the augite syenite unit of the 1.13 Ga alkaline to peralkaline Ilimaussaq intrusion in South Greenland contain olivine (Fa75-85) with micrometer-sized exsolution lamellae of intergrown augite and magnetite. The exsolved olivine contains only about 0.6 wt% CaO, whereas the reintegrated magmatic composition contained about 1.4% CaO corresponding to about 3 mol% of a larnite (La) component. The exsolved olivine bears no measurable Na, Al, or Ti, but the reintegrated magmatic olivine contained up to 0.4 wt% Na 2 O, 0.2 wt% Al 2 O 3 , and 0.3 wt% TiO 2 . Magmatic baddeleyite in the same samples indicates an extremely low silica activity of below 0.4 during the early stages of magmatic crystallization at > 900 °C and 1 kbar. According to the equilibrium Ca 2 SiO 4 (La in Ol) + SiO 2 = 2 CaSiO 3 (Wo in Cpx), the low silica activity is thought to be responsible for the unusually Ca-rich compositions of the primary olivines, which, in these samples, formed at 875 ± 15 °C and at oxygen fugacities between 1.2 and 2.3 log units below the QFM buffer. The composite exsolution lamellae formed during cooling due to an overstep of the schematic reaction Fa + Kst (Ca-rich component in olivine) + 1/3 O 2 = Hd + 2/3 Mt. Phase equilibria indicate temperatures between 300 and 750 °C and oxygen fugacities close to QFM for this exsolution process. The lamellae only developed in samples where the olivines were both especially high in Ca and presumably rapidly cooled (although not chilled!), as the lamella-bearing samples were collected at the outermost rim of the augite syenite unit and close to the top of the Ilimaussaq intrusion. Schreinemakers analysis shows that Ca-rich olivine of the kirschsteinite-monticellite solid solution series [Ca(Fe,Mg)SiO 4 ] would exsolve from Ca-bearing olivine of the fayalite-forsterite series only at unusually low silica activities, whereas at higher silica activities, the observed exsolutions of clinopyroxene with magnetite are more stable. Hence, both cooling history and evolution of intensive variables during cooling are important for the formation of olivine exsolution textures.

Single crystals of San Carlos olivine in contact with graphite were annealed at P = 1 bar, T = 1373 K, for studying the reaction of extraction of (Fe, Ni) metal. Scanning electron microscopy and transmission electron microscopy were performed on samples recovered after the experiments. Precipitates of (Fe, Ni) and thin amorphous layer of silica were identified, exclusively on the surface of the single crystals. Mass balance indicates that volatilization of Fe, Mg, and Si is negligible under these conditions. The reaction can be summarized as: Fe 2 SiO 4 in olivine + 2C graphite = 2Fe in metal + SiO 2 amorphous + 2CO in gas which occurs at the crystal surface without affecting the interior of the crystal, except for an Fe 2+ and Mg 2+ compositional profile in the olivine matrix. These chemical profiles are consistent with measured values of Fe 2+ -Mg 2+ interdiffusion coefficients, in agreement with the fact that Si and O are relatively immobile in olivine under such conditions. This study shows that annealing at relatively moderate temperature under reducing conditions can cause surface modifications and thus probably can strongly influence the surface evolution of planetary objects exposed directly to space environments (regoliths, surfaces of asteroids, or interplanetary dust particles).

Leverage analysis enables identification of reflections with the greatest influence on the estimate of each refined variable, and thus may be an important tool to improve standard structure-refinement procedures, especially in the case of minerals with complex composition. In this work, leverage analysis was used to investigate in detail the influence of each reflection in high-resolution X-ray diffraction data collected from olivine, a mineral often used to model order-disorder processes and to calculate temperatures of closure and cooling rates of host rocks. Particular attention was paid to the estimate of the scattering power at the cation sites, that are crucial for the above studies. Various procedures for data correction and refinement were also investigated, and the different possible choices were compared to choose the strategy that provides the best results. The presence of high-leverage weak reflections in olivine strongly suggests that systematic data truncation according to intensity threshold should be avoided. The estimates of the site-scatterings obtained under the different conditions tested are very close (always ≤3 σ); they are often smaller than those which may be obtained from electron-microprobe analysis under different experimental conditions or on inhomogeneous (zoned) crystals. Chemical data should thus not be routinely used to constrain the refinement procedure and/or to optimize final site-populations, provided that appropriate errors are given; on the other hand, they are valuable to appreciate the presence and the amounts of very minor, sometimes unexpected, substituents. Our tests show that precision better than 0.001 in site-occupancy determination (sometimes claimed in the literature) is probably not achieved.

A set of natural and heat-treated cordierite crystals has been analyzed using single-crystal X-ray diffraction, EMPA, SIMS, and 57Fe Mössbauer-spectroscopy. Structure determination of natural cordierite shows that the average size of the tetrahedrally coordinated T 1 1 cation increases systematically by about 0.01 Å as the Fe content of the solid solution decreases toward the Mg end-member. For Fe-rich compositions, the mean tetrahedral bond length T 1 1-O is close to 1.749 Å, a value expected for an AlO 4 tetrahedron in an alumosilicate framework structure. It is suggested that the structural dilation of the tetrahedral sites, which is driven by the decreasing average size of the octahedral cation, is sustained by substitution of the large cations Mg and Fe 2+ for Al. Mössbauer spectroscopy shows that up to 11% of Fe 2+ can be attributed to tetrahedral coordination in Mg-rich cordierite. Charge balance for the substitutions is either provided by introduction of Na + into vacant Ch2(0,0,0) channel sites or by substitution of an additional Al 3+ by Si 4+ . Tetrahedral site occupancies and corresponding Al,Si order parameters are calculated on the basis of a simple hard sphere model, based on refined cation-oxygen mean bond length, chemical composition, and Fe site-occupancy refinements. Almost complete Al,Si ordering among ring sites T 2 is encountered for most natural cordierite samples. Al,Si ordering among the T 1 tetrahedra is less pronounced for Fe-rich compositions. Partial disorder is attributable to Al/Si ratios in excess of 4/5. Orthorhombic shear strain is shown to correlate with the derived average order parameter, if corrections for strain contributions from size effects of the octahedral cation and from channel constituents are employed. Crystal-chemical similarities to other compounds and the implications of the minor cation substitutions for the properties of structural phase transitions in cordierite are discussed.

Enthalpies of drop solution in molten lead borate have been measured for a series of four zoisite and 12 clinozoisite samples, spanning both epidote solid solutions. Mixing within the zoisite series between 0.0 and 0.2 X Al₂Fe appears to be ideal. The enthalpy of drop solution varies between 493.0 ± 3.2 and 496.0 ± 3.6 kJ/mol with a slight increase with increasing Fe content, based on a linear fit to the data. The clinozoisite solid solution shows more complex behavior. The enthalpy of drop-solution for samples between 0.28 and 0.65 X Al₂Fe varies between 491.7 ± 4.2 and 500.7 ± 3.0 kJ/mol and shows significant scatter. For samples between 0.65 and 0.95 X Al₂Fe , the drop solution enthalpies increase abruptly, going from 494.3 ± 3.4 to 521.1 ± 5.4 kJ/mol. End-member drop solution values for Fe-free zoisite were estimated by the linear fit and for clinozoisite by assuming a fictive Fe-free end-member, taking into account the estimated energy of the monoclinic-orthorhombic transition, and by extrapolation of the Fe-rich data to X Al₂Fe = 1. Standard enthalpies of formation from the elements for Fe-free zoisite, Al-clinozoisite, and Fe-clinozoisite have been calculated based on the present data and data from the literature. The values are ΔH f,298.15 0,zoi = -6878.5 ± 6.8 kJ/mol; ΔH f,298.15 0,Al-czoi = -6882.5 ± 6.9 kJ/mol; ΔH f,298.15 0,Fe-czo = -6461.9 ± 6.8 kJ/mol. The overall trend in the drop solution data reflects an apparent positive enthalpy of mixing for the clinozoisite series, consistent with the presence of a miscibility gap in the series between about 0.5 and 0.8 X Al₂Fe . An attempt to fit the highly asymmetric enthalpy of mixing data with a two-parameter asymmetric Margules mixing model, however, failed to yield realistic values for the interaction parameters and the solvus closure temperature. The significant scatter in the data, between X Al₂Fe = 0.0 and 0.65, may be due to short-range ordering, suggested by Fehr and Heuss-Assbichler (1997), or could be due to different states of metastable disorder in the samples. The steep increase in enthalpy going from X Al₂Fe = 0.65 to 1.0 the may be due to the effect of possible increasing Fe 3+ -Al 3+ disorder among the M1 and M3 octahedral sites. The break in slope of the enthalpy data occurs at or near the composition where the onset of Fe 3+ disorder had been observed previously in clinozoisite

Lead speciation was determined in a soil developed on a geochemical anomaly arising from a Pb-Zn stratabound deposit in Largentière (Ardèche, France). This geological setting offers the opportunity to determine the preferred form(s) of Pb following soil formation on this unique anomaly. In the soil profile studied, Pb concentrates in the B-horizon (2055 mg/kg Pb) relative to both the A- (1330 mg/kg Pb) and C- (1874 mg/kg Pb) horizons. Plumbogummite (PbAl 3 (PO 4 ) 2 (OH) 5 ·H 2 O) is the main host of Pb in the soil profile. Pb also appears to be associated with Mn-(hydr)oxides, as shown by micro-analyses (EMPA, SEM-EDS, and μ-SXRF), in the form of inner-sphere Pb 2+ complexes, as suggested by Pb L III -edge EXAFS spectroscopy. Linear least-squares fitting of background-subtracted, k3-weighted Pb L III -edge EXAFS functions derived from bulk soil samples was carried out using Pb L III -EXAFS spectra of 22 Pb-containing model compounds. Quantitative assessment of Pb speciation revealed that, whereas plumbogummite is the most abundant Pb phase in the soil profile, Pb 2+ -Mn-(hydr)oxide surface complexes are gradually replaced by Pb 2+ -surface complexes with other phases, possibly Pb 2+ -organic complexes, upward in the soil profile. The presence of large amounts of Pb-phosphate in the Largentière soil suggests that low solubility phosphates may be important long-term hosts of Pb in Pb-contaminated soils that have sufficiently high phosphorous activities to cause formation of these phases.

To re-investigate “fundamental particles” and “interparticle diffraction” proposed by Nadeau et al. (1984a, 1984b, 1984c), we made physical mixtures of smectite and rectorite and examined them by X-ray diffraction analysis and transmission electron microscopy. Smectite and rectorite consisting of at least 10 and 23 silicate layers in their parent rocks, respectively, reduce their thickness by dispersion in water and a fractionation procedure that releases thin silicate layers from the original crystals. The <0.1 μm fractions of the smectites are composed of mostly two to three silicate layers that are not fundamental particles. The mixtures of smectite and rectorite of 0.5-2 μm fractions are mere physical mixtures and do not show interparticle diffraction. In contrast, the mixtures of smectite and rectorite of <0.1 μm fractions have interstratified structures of a segregation type and show interparticle diffraction. However, interparticle diffraction does not occur for synthetic hectorite with aspect ratios smaller than those of the above smectites of the <0.1 μm fractions. These results strongly suggest that interparticle diffraction occurs, and that interstratified structures are formed by mixing rectorite and smectite physically and preparing the oriented specimens when the particle size in the a-b plane and the particle thickness are appropriate. However, such mixtures do not reflect the actual characteristics of natural smectite and illite. A rectorite particle of about 30 nm thickness consists of sub-particles of two or more rectorite units, or four or more 2:1 layers that are nearly perfectly coherent with one another, and the sub-particles are turbostratically stacked. Our results suggest that natural, interstratified illite/smectite are not accumulations of fundamental particles but represent their structural sequences when formed.

Changes produced after hydrothermal treatments at 400 °C and at different pressures in a set of Lu-saturated smectites have been compared with those occurring in a mixture of 2SiO 2 :1Lu 2 O 3 . The effect of the mineralogical composition of the smectite on the reactivity has been analyzed. The growth of a new crystalline phase, Lu 2 Si 2 O 7 (detected by means of X-ray diffraction), at a different pressure for each smectite, has allowed us to establish a reactivity order among the samples. All the samples were much more reactive than the mixture of oxides. The reactivity of the smectites was modulated by their different mineralogical compositions. First, smectites having Al in the tetrahedral sheet were the most reactive. This finding is interpreted as due to a strong electrostatic interaction created by the Al that activates neighboring Si for the reaction. Second, a higher reactivity was observed for the smectites with total occupancy of the octahedral sheet, due to the simultaneous disruption of the tetrahedral-octahedral shared oxygen layer and the diffusion of octahedral cations into the interlayer space. Third, the layer charge of the smectite does not induce any variation in the reactivity for the set of samples analyzed.

The short-range order of a set of Lu-saturated smectites, modified by hydrothermal treatment at 400 °C, which give rise to a new crystalline phase (Lu 2 Si 2 O 7 ), has been analyzed. In particular, the modifications undergone by the tetrahedral Si atoms of the lattice and the possible diffusions of O-T layer ions to other structural positions were investigated. Single-pulse, magic-angle-spinning nuclear magnetic resonance (MAS-NMR) measurements from the 1 H, 7 Li, 27 Al, and 29 Si nuclei, as well as two-dimensional 1 H- 27 Al cross-polarization MAS-NMR experiments, have been made on both the initial and the hydrothermally treated samples. The results show that (1) the formation of the Lu 2 Si 2 O 7 takes place from Si atoms surrounded by Al tetrahedra in the samples that contain tetrahedral Al; (2) there is a diffusion of octahedral cations to the interlayer positions during the hydrothermal treatments, compatible with the X-ray diffraction data reported for this reaction; and (3) the octahedral Al migrates to the vacant tetrahedral positions that appear during the formation of the new Lu 2 Si 2 O 7 phase. A clear relationship between each sample composition and its reactivity is defined.

Self-diffusion coefficients of water molecules (H 2 O) in Na-montmorillonite gel were measured as a function of water fraction (54.8 to 100 wt%) and temperature (30.4 to 60.0 °C) using proton nuclear magnetic resonance (NMR). Spin-echo pulse sequences with magnetic field gradient pulses for the diffusion measurement were applied to the montmorillonite gel at the Larmor frequency of 20 MHz. The self-diffusion coefficient, D, of H 2 O in the clay gel is expressed phenomenologically by ln(D/D 0 ) = 1.77{exp[0.0798(w − 100)] − 1}, where D 0 is the water diffusivity in bulk water and w is the water fraction of the gel (wt%). The data for w > 84.7 wt% can be explained by a theoretical diffusion model for the randomly distributed clay grains. The activation energy of the water diffusivity in the montmorillonite gel was nearly equal to that in bulk water, so the normalized diffusivity, D/D 0 , obeys the temperature-independent master curve. The transition from the free diffusion to the restricted diffusion was not observed for gradient pulse intervals ranging from 5 to 120 ms. This indicates that the average pore size of the gel is much smaller than a few tens of micrometers, so the random walk trajectory of water molecules in the gel is geometrically restricted by the packing of clay mineral grains. Water diffusivity higher than that of the present NMR study was found by computer simulations and neutron scattering experiments in which effects of bound water were considered but those of the tortuosity of the grain packing were neglected. Thus, the predominant factor controlling the diffusivity in the NMR experiments is not the bound water near the clay surface but the geometrical tortuosity of the packing of clay mineral grains.

Both X-ray absorption and diffraction techniques were used to study the structural environment and oxidation state of Mn in goethite-groutite solid solutions, α-Mn x Fe 1-x OOH, with x Mn ≤ 0.47. Rietveld refinement of X-ray diffraction (XRD) data was employed to investigate the statistical long-range structure. The results suggest that increasing x Mn leads to a gradual elongation of Fe and Mn occupied octahedra which, in turn, causes a gradual increase of the lattice parameter a and a gradual decrease of b and c in line with Vegard’s law. X-ray absorption fine structure (XAFS) spectra at the MnKα and FeKα edges revealed, however, that the local structure around Fe remains goethite-like for x Mn ≤ 0.47, while the local structure around Mn is goethite-like for x Mn ≤ 0.13, but groutite-like for higher x Mn . The spectral observations were confirmed by XAFS-derived metal distances showing smaller changes around Fe and larger changes around Mn as compared with those determined by XRD. Therefore, the XAFS results indicate formation of groutite-like clusters in the goethite host structure for x Mn > 0.13, which remain undetected by XRD. The first prominent resonance peak in the X-ray absorption near-edge spectra (XANES) of the Mn goethites was 17.2 to 17.8 eV above the Fermi level of Mn (6539 eV), in line with that of Mn 3+ reference compounds, and well separated from that of Mn 2+ and Mn 4+ compounds. Therefore, Mn in goethite is dominantly trivalent regardless of whether the samples were derived from Mn 2+ or Mn 3+ solutions. This may indicate a catalytic oxidation of Mn 2+ during goethite crystal growth similar to that found at the surface of Mn oxides.

Henritermierite, space group I4 1 /acd, at 293 K a = 12.489(1), c = 11.909(1) Å, Z = 8, with close to end-member composition (Ca 2.98 Na 0.01 Mg 0.01 ) VIII (Mn 1.95 Fe 0.01 Al 0.04 ) VI [SiO 4 ] 2.07 [H 4 O 4 ] 0.93 from the N’Chwaning II mine at the Kalahari manganese fields, Republic of South Africa, has been studied by single-crystal X-ray diffraction at 100 and 293 K at ambient pressure and up to 8.7 GPa in a diamond-anvil cell at 293 K. Polarized FTIR spectroscopy at 80 and 293 K was also performed. The Mn 3+ O 6 octahedra display a tetragonally elongated type of Jahn-Teller distortion where the oxygen atoms of the elongated O-Mn-O axis (Mn-O: 2.2 Å) are moderately hydrogen bonded (O-H…O: 2.76 Å) to the H 4 O 4 tetrahedra, which replace 1/3 of SiO 4 tetrahedra in an ordered fashion. Thus Jahn-Teller distortion and H 4 O 4 arrangement are coupled and both are responsible for the tetragonal bulk symmetry. The H 4 O 4 tetrahedra have a center-to-O distance of 1.98 Å and the H atoms are slightly above the tetrahedral faces as similarly observed in the synthetic katoite end-member, Ca 3 Al 2 [H 4 O 4 ] 3 . However, in henritermierite the O-H…O hydrogen bond is considerably bent (ca. 131°) and gives rise to an OH stretching mode at 3432(5) cm -1 . Additional, though weak, IR absorptions at 3508(2) and 3553(2) cm -1 may be due to more remote hydrogen-bond acceptors (O-H…O: 3.29 Å) within the H 4 O 4 tetrahedra. Compressibility data for a third-order Birch-Murnaghan equation of state yield a bulk modulus of K 0 = 97.9(9) GPa with a pressure derivative of K' = 5.3(3). The axial compressibilities indicate a pronounced compressional anisotropy which is explained by the orientation of the elongated axes of the Jahn-Teller distorted MnO6 octahedra along the slightly more compressible [100] tetr directions compared to the c-axis. The crystal structure was refined at a pressure of 8.6 GPa. The MnO 6 octahedra were observed to show anisotropic compression towards a more isometric shape. Calculated spontaneous strain reveals a trend towards a weaker tetragonal distortion. If henritermierite is heated above 800 K in air it dehydrates and Mn 3+ is partially oxidized to Mn 4+ . This topotactic transformation leads to a new garnet-like phase of Ia3d symmetry with a = 12.12 Å and of Ca 3 Mn 2.26 O 2.32 [SiO 4 ] 2.42 composition in which instead of H 4 O 4 tetrahedra a new disordered octahedral site is occupied by Mn

High-resolution single-crystal X-ray diffraction intensities recorded with a charge-coupled device detector at low temperature (100 K) were used to derive the electron density distribution in three Al 2 SiO 5 polymorphs: andalusite, sillimanite, and kyanite. The 27 Al nuclear quadrupole coupling tensors were estimated from both the internal polarization contribution of the Al electron density and that of the lattice with electronic multipoles up to octupoles for oxygen atoms and hexadecapoles for Si and Al. Based on new estimations of the 27 Al Sternheimer shielding factor, R and the antishielding factor, γ, a close agreement was achieved between the quadrupole coupling tensors at the various Al-sites derived from the X-ray diffraction data for all three polymorphs of Al 2 SiO 5 with those determined previously from single-crystal 27 Al NMR spectra.

Static and 23 Na MAS and 1 H MAS NMR experiments at temperatures down to 100 K as well as high temperature static and 23 Na MAS NMR experiments up to 873 K were carried out to study the dynamics of the sodium cation in nitrate cancrinite. In addition, X-ray powder data were recorded at 296 K and 673 K to obtain information about long-range order changes in the structure. Low temperature 23 Na NMR experiments show almost no change in signal, whereas the 1 H MAS NMR spectra show the freezing of cage water motion. The two different crystallographic sodium sites cannot be distinguished in the 23 Na MAS NMR spectra at low and room temperature. In contrast, two 23 Na NMR signals exist at high temperatures above 673 K, a quadrupolar pattern and a Lorentzian signal. Rietveld structure refinements of powders show no change in the local environment of sodium. The strong quadrupolar coupling results from the loss of cage water and the near planar O atom coordination of Na in the ε-cage, while the interaction of Na + in the channel remains similar. Thus the change in quadrupolar interaction is mainly caused by the absence of electrostatic shielding of cage water after dehydration

The crystal structure of OD-chondrodite [Mg 5 Si 2 O 8 (OD) 2 , P2 1 /b (a unique), a = 4.74711(5), b = 10.34888(16), c = 7.90228(13) Å, α = 108.678(1)°] was refined to wR p = 0.0218, χ 2 = 3.545 at ambient conditions using time-of-flight neutron powder data. The disordered H model proposed for OH-chondrodite on the basis of single-crystal X-ray data is confirmed. The occupations of the D1 and D2 sites are, respectively, 0.52(1) and 0.48(1). The long O5-D1 [1.076(4) Å] and O5-D2 [1.111(4) Å] bond lengths, which are two of the longest O-H(D) bonds observed in mineral structures, reflect the positional disorder of the O5 atom in the unshared OH-OH edge. Both D1 [1.968(4), 2.489(4) Å] and D2 [2.149(4), 2.251(4) Å] atoms are involved in two hydrogen bonds. A re-examination of the origin of positive OH frequency shifts in both F-bearing and OH-chondrodite at high pressure is warranted in view of the crystallographic data.

The structure of synthetic hydroxyl-clinohumite [Mg 7 Si 4 O 14 ·2Mg(OH)O] was refined for a deuterated sample using powder neutron diffraction data and the Rietveld technique [(P2 1 /b; Z = 2; a = 4.7488(1) Å; b = 10.2875(2) Å; c = 13.6967(3) Å; α = 100.63(1)°; V = 657.65(2) Å 3 )]. The H atoms of the hydroxyl groups are disordered over two positions with an occupancy of approximately 0.5. This removes the possibility of a strong H-H interaction that was thought to destabilize the structure. The limited occurrence of hydroxyl-clinohumite in nature is the result of compositional rather than crystal chemical constraints

The structural response of the natural zeolite yugawaralite (CaAl 2 Si 6 O 16 ·4H 2 O) upon thermally induced dehydration has been studied by Rietveld analysis of temperature-resolved powder diffraction data collected in situ in the temperature range 315-791 K using synchrotron radiation. The room-temperature monoclinic structure [Pc, a = 6.73200(9), b = 14.0157(2), c = 10.0607(1) Å, β = 111.189(1) °, Z = 2, at 315 K] has the Ca cations in the channels coordinated to four framework O atoms and to four water molecules, with two of the water sites (OW1 and OW4) showing positional disorder progressively disappearing as the dehydration proceeds. The yugawaralite structure reacts to the release of water molecules with small changes in the Ca-O bond distances and minor distortions of the tetrahedral framework up to about 695 K. Above this temperature the Ca coordination falls below 7 (four framework O atoms and three water molecules) and a major rearrangement in the cation coordination takes place, causing a first order phase transition involving both a large decrease in the cell volume and the change in the space group symmetry. A satisfactory structure model for the high-temperature phase stable in the range 695-791 K could not be obtained because of the complexity of the structure. A model approximately describing the average structure [Pn, a = 12.703(1), b = 13.067(1), c = 9.839(1) Å, β = 110.894(9), Z = 4, at 749 K] has been used to follow the temperature evolution of the cell parameters in the measured temperature range. This model involves a sixfold coordination of the Ca cations (five framework O atoms and one water molecule). There is no indication of significant structure changes before collapse, likely occurring when the last water molecule is expelled from the structure and the cation coordination drops below 6, as observed in other Ca-rich zeolites (i.e., laumontite, scolecite, mesolite).