Volume 94, Issue 5-6

The closure temperature, T C , of the intracrystalline ordering of Mg-hercynite is estimated with a comparative crystal-chemical approach. The single crystals were selected from two distinct geological environments that represent extremely different cooling rates. The fast cooled setting refers to anatectic metapelitic enclaves that occur in the high-K calc-alkaline lavas of the Neogene Volcanic Province of SE Spain. The slow cooled setting refers to metabauxite from the Anga metamorphic complex, Lake Baikal. Parameters sensitive to T C include the oxygen fractional coordinate (u) and the inversion parameter (i). Experimental equilibration data on the spinel and hercynite end-members and on their solid solution are fitted to equations where TC is given as a function of the hercynite content (Hc) of the solid solution and of u or i. The unavoidable simplifications made in this empirical approach are discussed. A reasonable value for T C , ~400 °C, was obtained for the slow cooled metamorphic hercynite from the oxygen fractional coordinates. In contrast, an unreasonably high value of T C , ~600 °C, was obtained from the inversion parameters. In the case of the fast cooled anatectic samples, TC calculated from the two structural parameters are comparable; the five crystals show a range in the calculated values for T C over ~250 °C, from ~700 to ~950 °C, which is reasonable considering the known diversity of cooling rates exhibited by their volcanic host-rocks.

Arsenate (As 5+ ) substitution in the hydroxylapatite structure was examined using a combination of crystallographic and spectroscopic techniques. Samples of hydroxylapatite, the As 5+ -substituted analog (synthetic johnbaumite), and five intermediate compositions were synthesized from solution. Synchrotron X-ray diffraction data show that all samples are single-phase, confirming complete substitution. No evidence is found for lowering of symmetry below P6 3 /m. Rietveld structure refinements show progressive expansion of the unit cell with increasing As substitution, which can be accounted for primarily by an average expansion of the tetrahedral site. Sizes of Ca polyhedra show little variation as a result of As substitution. NMR results show no evidence for local clustering of PO 4 tetrahedra. EXAFS confirms that the size of As-centered tetrahedra remains constant across the solid-solution series.

Kintoreite, PbFe 3 H 0.94 [(PO 4 ) 0.97 (SO 4 ) 0.03 ] 2 (OH) 6 , and zincian kintoreite, PbZn 0.3 Fe 3 H 0.24 [(PO 4 ) 0.54 (SO 4 ) 0.08 (AsO 4 ) 0.38 ] 2 (OH) 6 , have rhombohedral symmetry, space group R3m, with hexagonal cell parameters a = 7.2963(5) Å, c = 16.8491(5) Å, and a = 7.3789(3) Å, c = 16.8552(7) Å, respectively. The structures have been refined using single-crystal X-ray data to R 1 = 0.030 for 374 observed reflections and R 1 = 0.035 for 399 observed reflections, respectively. The structures of both minerals comprise rhombohedral stacking of (001) composite layers of corner-shared octahedra and tetrahedra with Pb atoms occupying icosahedral sites between the layers, as in the alunite-type structure. The cornerconnected octahedra form three-membered and six-membered rings as in hexagonal tungsten bronzes. The structure of zincian kintoreite differs from other alunite-type structures in having partial occupation, by Zn, of new sites within the six-membered rings in the octahedral layers. The Zn is displaced to an off-center position in the hexagonal ring, where it assumes fivefold trigonal-bipyramidal coordination, to three of the hydroxyl anions forming the ring, and to the apical O anions of the XO 4 tetrahedra on opposite sides of the ring. The different structural modes of Zn incorporation into SO 4 -dominant and (P,As)O 4 -dominant members of A 2+ B 3 3+ (XO 4 ) 2 (OH) 6 alunite-type minerals are discussed in terms of the different charge-compensation mechanisms involved.

Precession electron diffraction is used to characterize a (021) twin observed in coesite. Due to the quasi-hexagonal dimensions of coesite (monoclinic space group C12/c1 with β = 120.34°), indexing of conventional spot patterns is ambiguous and the twin law determination is impossible. With precession, spot intensities enable the absolute indexing of the patterns. The method we used is based on the analysis of the departure from hexagonal symmetry. This ensures that all possible pseudo-equivalent monoclinic orientations are taken into account for the indexing. The orientation relationships between adjacent parts of the twin are then characterized. The twin is described as a mirror along (021), which is consistent with original descriptions of twinning in synthetic coesite and a previous characterization performed using large-angle convergent-beam electron diffraction (LACBED).

Molecular dynamics simulations have been used to study the structure, equation of state (EOS), self-diffusion, and shear viscosity of molten Mg 2 SiO 4 for pressures and temperatures in the range 2.5-110 GPa and 2100-5060 K, respectively. The transferable pair-potential parameters of Matsui (1998) for the system Na 2 O-CaO-MgO-Al 2 O-SiO 2 have been used accounting for Coulomb, Born, and van der Waals forces. Simulations have been carried out in the microcanonical (NEV) ensemble at 63 state points along 12 isochores spanning the density range 2754-4500 kg/m 3 . Thermodynamic properties including the isochoric heat capacity, isobaric expansivity, isothermal compressibility, thermal pressure, and the Grüneisen parameter (γ) are computed directly from MD results. A density crossover between molten Mg 2 SiO 4 and forsterite crystals occurs at ~15 GPa at 2100 K. We find the Grüneisen parameter to be a function of temperature (T), increasing with increasing T at low density (ρ < 3400 kg/m 3 ) but decreasing as T rises at high density (ρ > 3400 kg/m 3 ); hence, the integrated form of the Mie-Grüneisen EOS is only approximately valid for liquid Mg 2 SiO 4 since γ varies by ~20% over the T range along an isochore. Radial distribution functions for all atoms around all other atoms were used to generate coordination statistics as a function of pressure (P) and T. Oxygen about Si coordination increases from fourfold coordination at low pressure to sixfold at higher pressure; the abundance of distorted trigonal bipyramidal fivefold polyhedra, Si(V) maximizes at 30 GPa at 3500 K. Interestingly, O about O increases to a maximum of 13 at low P before decreasing with increasing pressure to ~10. The mean coordination number (CN) of Si around oxygen increases from 1.2 to 1.5 consistent with an increasing abundance of Si 2 O 7 dimers as pressure increases. Self-diffusion of Mg, Si, and O was calculated at each state point giving activation energies of 67, 79, and 76 kJ/mol and activation volumes of 1.42, 1.10, and 1.32 cm 3 /mol, respectively. Shear viscosity of the liquid calculated at 12 state points using the Green-Kubo formulation provides an excellent Arrhenian fit. Viscosity varies by a factor of ~20 (1.5 × 10 -3 Pa s to 0.03 Pa s) from 1 to 100 GPa. The validity of the Stokes-Einstein and Eyring expressions for atom mobility and shear viscosity is examined in detail. Characteristic lengths for atom mobility are consistent with ionic radii to within a factor of ~1.5 -2 for all atoms. An equation of state and thermodynamic model for Mg 2 SiO 4 liquid is developed consistent with the fundamental measure functional theory of Rosenfeld and Tarazona (1998). Our model reproduces the E-P-V-T relations and the derived thermodynamic properties obtained from the MD simulations to within the reported uncertainty.

Åkermanite (Ca 2 MgSi 2 O 7 ) and gehlenite (Ca 2 Al 2 SiO 7 ) have been studied at high pressure by synchrotron radiation powder and single-crystal diffraction up to 30 GPa. At about 2 GPa, the incommensurately modulated structure (IC) transforms to a normal structure (N). The bulk modulus for the N structure, fitted with a Birch Murnhagan EoS on powder data, is 93.5(5) GPa. The compressibility is anisotropic, and it is greater along the c axis, in the direction perpendicular to the tetrahedral layers of the structure. Above 15 GPa, a phase transition is observed, marked by a discontinuity in the elastic behavior and a small change in intensity and in the full-width at half maximum (FWHM) of the powder diffraction peaks. The diffraction patterns are indexed with respect to tetragonal cell of the Nmelilite structure up to 30 GPa. A hysteresis in the elastic behavior is observed during decompression. In contrast, single-crystal data show a new monoclinic phase appearing above 15 GPa. The unit-cell parameters are a = 8.82(1) Å, b = 7.34(1) Å, c = 9.13(1) Å, β = 115.1(2)°. This unit cell is similar to that of Ca 2 ZnGe 1.25 Si 0.75 O 7 reported in the literature. A refinement using the corresponding model in space group P2 1 /n fits the single-crystal data with a reasonable R Bragg = 15%, considering that the crystal is twinned and the mosaicity is large. Gehlenite has a higher bulk modulus, 106.1(4) GPa, than does åkermanite. The compressibility is anisotropic, and the behavior is similar to that of åkermanite, but the presence of Al in tetrahedral sites decreases the compressibility parallel to the (001) plane. The structure of gehlenite is stable up to 25 GPa, when a phase transition occurs.

In New Caledonia, the weathering of ultamafic rocks under a tropical climate has led to the residual accumulation of trace elements in lateritic soils widely dominated by Fe-oxyhydroxides. The speciation of trace elements, such as Cr, Ni, and Co, in these Oxisols remains a major subject of interest regarding mining and environmental issues. We have assessed the speciation of chromium in the upper part of an Oxisol, by combining bulk and spatially resolved chemical analyses (EPMA and SEM-EDS) with synchrotron-based spectroscopic data (EXAFS and XANES). EPMA indicates that the main hosts for chromium in the bedrock sample are the silicates forsterite, enstatite, and lizardite. Hosting of chromium in these easily weatherable mineral species could lead to a significant loss of this element upon weathering. However, total analyses of major elements indicate only a slight depletion of Cr, together with an immobility of Fe and Al and drastic losses of Si and Mg, after the weathering of the bedrock. Such a low mobility of chromium is likely related to its significant incorporation in goethite and hematite formed after the weathering of Fe 2+ -bearing primary silicates. This efficiency of secondary Fe-oxyhydroxides at immobilizing chromium is demonstrated by quantitative analysis of EXAFS data that indicates that these mineral species host between 67 and 75 wt% of total Cr (compared to the 18 to 22 wt% of total Cr hosted by chromite). In addition, SEM observation and SEM-EDS analyses performed on the Oxisol samples also show some evidence for chemical weathering of chromite. Chromite could then represent a past and/or present source of chromium upon extended tropical weathering of the studied Oxisol, rather than a stable host. These results emphasize the importance of secondary Fe-oxyhydroxides, compared to Cr-spinels, on chromium hosting in Oxisols developed upon tropical weathering of ultramafic rocks. Although the trapping mechanism of chromium mainly corresponds to incorporation within the structural network of goethite and hematite, sorption reactions at the surface of these mineral species could also be involved in such a process. In addition, considering their potential oxidative reactivity that can generate Cr 6+ or enhance the chemical weathering of chromite, the occurrence of Mn oxides could significantly modify the behavior of chromium upon weathering. These considerations indicate that further studies are needed to assess the actual potential of chromium release from Oxisols developed upon weathering of ultramafic rocks under a tropical climate.

The new mineral meurigite-Na, [Na(H 2 O) 2.5 ][Fe 3+ 8 (PO 4 ) 6 (OH) 7 (H 2 O) 4 ], is monoclinic with space group C2/c and cell parameters a = 28.835(2), b = 5.1848(4), c = 19.484(1) Å, β = 106.983(6)°, V = 2785.8(2) Å 3 , and Z = 4. It is the Na analog of meurigite, which is now named meurigite-K. The type locality of meurigite-Na is the Silver Coin mine, Valmy, Iron Point district, Nevada, where it occurs as radial sprays of cream-colored, flattened fibers (thin laths) to 0.4 mm in length. At the Silver Coin mine, meurigite-Na occurs very late in a paragenetic sequence that includes (in approximate order from early to late) quartz, barite, apatite-(CaF), goethite, rockbridgeite, cacoxenite, alunite, wardite, turquoise/chalcosiderite, leucophosphite, lipscombite/zinclipscombite, kidwellite, strengite/ variscite, crandallite/perhamite, and jarosite. Meurigite-Na also occurs at Tom’s quarry and Moculta quarry in South Australia, Australia, Lake Boga quarry and probably Rixon’s Sandstone quarry in Victoria, Australia, and at an unnamed pegmatite prospect near Linopolis, Minas Gerais, Brazil. The streak is white, the luster is silky, and the Mohs hardness is about 3. The measured density is 2.94(2) g/cm 3 and the calculated density is 2.954 g/cm 3 . Optical properties: biaxial (-), α = 1.740(3), β = 1.759(3), γ = 1.763(3), 2V meas = 50(10)°; X ≅ c, Z = b. The strongest powder X-ray diffraction lines are [d(hkl)I]: 13.8(200)20, 9.35(002)100, 4.843(111)20, 3.206(7̄̄̄̅12)40, 3.107(7̄̄̄̅13)30, 2.971(513)15, and 2.593(116)15. Meurigite-Na is insoluble in concentrated HCl. Evaluation of powder XRD and chemical-analytical data in light of the structure determination for meurigite-K leads to the conclusion that type phosphofibrite is isostructural with meurigite and represents the alkali-deficient member of a series with meurigite-K.

Structural investigations of several minerals belonging to the calaverite group with composition Au 1-x Ag x Te 2 (x = 0.00, 0.02, 0.05, 0.09, 0.19, and 0.33) indicate that Ag is randomly distributed on the Au sites. This suppresses the valence fluctuation of Au and, therefore, the structure modulations. The results are compared with the previously published incommensurately modulated structure of calaverite, Au 0.9 Ag 0.1 Te 2 , which is characterized by valence fluctuations of Au reinforced by an ordered distribution of Ag. The (3+1)-dimensional calaverite structure type is able to reproduce both (3+1)D and 3D related structures with the general formula AB 2 (A = Au, Ag, Cu, Nb, Ta; B = Te).

Li-rich zincostaurolite occurs as millimeter-long crystals at the marble footwall of a meta-karstbauxite on eastern Samos. The Samos rocks have been metamorphosed during an early Alpine high-P, low-T metamorphism (M1) followed by a late Alpine greenschist-grade overprint (M2). Textures and mineral chemistry indicate that staurolite formed from gahnite, cookeite, and pyrophyllite during the early M1 stage. Staurolite crystals show growth zoning with cores enriched in Zn. Concentrations of Fe, Mg, Co, and, to a minor extent, Li increase toward the rims. Hydrogen concentrations were analyzed by SIMS. They are significantly higher in cores (up to 5.97 atoms H per 48 O) compared to rims (3.9 to 4.5 atoms H) and clearly negatively correlated with Al. Synchroton-light polarized FTIR spectra on oriented FIB-prepared foils show the same zonation effect, the absolute hydrogen concentrations being systematically lower by about 25%. The discrepancy is caused by sub-micrometer scale hydrogen loss at the crystal surface during FIB-thinning. This staurolite is unique as from the three available hydrogen sites the H3 site has the highest occupation ever observed, whereas the H2 site is not occupied. This is probably due to the high Li content. The zonation in hydrogen is interpreted as reflecting the two-stage growth. M1-staurolite that formed a low T of about 400-450 °C and high P of >1.5 GPa incorporated nearly the maximum amount of hydrogen allowed by the staurolite structure (6 H pfu) and was subsequently overgrown and marginally replaced during the M2 stage by less hydrous, Fe-Co richer staurolite. Hydrogen zoning in staurolite is facilitated by the sensitivity of its structure to changing P-T conditions. Water in staurolite is maximized at high P and low T. Cores of staurolite from Samos represent the most hydrous staurolite compositions reported to date.

Dmitryivanovite (CaAl 2 O 4 ) is a newly described, calcium aluminum oxide from the Northwest Africa 470 (NWA470) CH3 chondrite (Ivanova et al. 2002). NWA470 contains abundant small Ca,Alrich inclusions (CAIs), and dmitryivanovite, whose composition is close to stoichiometric CaAl 2 O 4 [Ca 1.000 (Al 1.993 Si 0.003 Ti 0.002 ) 1.998 O 4 ], was found in one of these CAIs. It occurs as ~10 μm subhedral grains intergrown with grossite (CaAl 4 O 7 ), perovskite, and melilite. Electron backscatter diffraction (EBSD) analysis revealed that dmitryivanovite is a high-pressure polymorph of CaAl 2 O 4 (a = 7.95, b = 8.62, c = 10.25 Å, β = 93.1°, space group P2 1 /c, and Z = 12). Dmitryivanovite is the third phase to be described from nature in the binary system of CaO-Al 2 O 3 , the other two being hibonite (CaAl 12 O 19 ) and grossite (CaAl 4 O7)-all are found in CAIs. The presence of CaAl 2 O 4 in NWA470 suggests a local elevated dust/gas ratio in the solar nebula. The phase diagram of CaAl 2 O 4 shows that ~2 GPa is required to stabilize the high-pressure CaAl 2 O 4 polymorph at 1327 °C, above which CaAl 2 O 4 condenses from the solar nebula. Because it is unlikely that the solar nebula ever had such a high total gas pressure, it appears more probable that condensation of the low-pressure polymorph occurred in the solar nebula with an enhanced dust-to-gas ratio and that subsequently the high-pressure polymorph was produced by shock metamorphism, most likely after the CaAl 2 O 4 -bearing CAI was incorporated into the NWA470 parent asteroid.

The crystal structure of hydrous pure magnesium forsterite (Mg 2 SiO 4 ) containing 8900 ppmw H 2 O, synthesized at 12 GPa and 1250 °C, has been refined. The major hydration mechanism appears to be M1 cation vacancy with protonation of the O1-O2 octahedral edge of M1. Raman spectra up to 41 GPa show strong coupling between the two Ag modes (824.4 and 856.2 cm -1 ). Mode Grüneisen parameters γ i related to Mg(2)O 6 translation mode decrease relative to anhydrous forsterite. This is attributed to the fact that the M2 site shows full occupancy compared to the M1 site, and the vacancy predominantly occurs at the M1 site. Pressure dependencies of four OH vibrations in the region 3548-3615 cm -1 suggest that positional ordering of hydrogen ion (proton) takes place with increasing pressure. The OH mode at 3615 cm -1 shows complex response as a function of pressure: the (∂ν/∂P) slope changes from -0.43 below 10 GPa to 1.97 between 10 and 20 GPa, and does not show clear pressure dependence above 20 GPa. Single-crystal X-ray data were used to assign the OH band to the structural sites, O1 and O2. The proton is closer to O2 than O1. The distance between the proton and O2 is 0.96 Å. In view of the empirical relation between OH-stretching frequencies and O···O distances (Libowitzky 1999), the pressure dependence of the OH mode (3615 cm-1) is well correlated with the O1-O2 distance and the degree of hydrogen bonding.

The geothermobarometric potential of tourmaline has been assessed by investigating element exchange among tourmaline and coexisting minerals in metamorphosed pelites and graywackes, and in experimental exchange between tourmaline and biotite. In the natural samples, a temperature dependence of tourmaline Mg-Fe exchange with biotite, staurolite, garnet, chlorite, and muscovite, and Ca-Na exchange with plagioclase is observed. Equilibrium calculations for the complete mineral assemblage show that tourmaline is in compositional equilibrium with all coexisting phases, which would allow for an internally consistent set of thermometers among all these phases to be defined. However, a prohibitively large spread is present in the K D vs. T relations. This is not caused by analytical effects, compositional zoning, or disequilibrium among the minerals. The experimental results show that it is the result of inter-site partitioning of elements over the Y and Z octahedral sites of tourmaline. Variations in the element distribution over these sites, their relative participation in the exchange and differences in the temperature dependence of exchange with each site, strongly affects the K D vs. T relation observed, with the slope actually changing sign depending on the elements residing at each site. Non-ideal interactions among the elements at each site will also affect this, and furthermore link every exchange to the bulk tourmaline composition, and hence the element mobility in the rock. The promising potential of tourmaline geothermobarometry can therefore not be fulfilled until effects of inter-site partitioning and non-ideal interactions are known.

Powder ZnAl 2 O 4 (gahnite) samples doped with 0-100 at% Co were obtained by a sol-gel technique. X-ray powder diffraction was used to characterize the samples. Gahnite samples are cubic with the normal spinel structure, space group Fd3̄m. Cobalt doping caused a nonuniform increase of unit-cell parameter. The structure of the gahnite samples was refined by the Rietveld method. The location of Co 2+ was determined by EPR spectroscopy. Cobalt doping of gahnite induces the inverse spinel structure at only 4 at% Co, and the inversion parameter increases with Co 2+ doping level. Metal-oxide distances in the (Al,Co)O 6 octahedra dominantly influence the unit-cell parameter of Co-doped gahnite

This study is focused on the behavior of ternary SiO 2 -Na 2 O-B 2 O 3 borosilicate glasses at temperatures between 298 and 1800 K. Unpolarized Raman spectra were measured up to high temperature. SiO 2 -Na 2 O-B 2 O 3 glass samples were prepared with different values of the ratio R = [Na 2 O]/[B 2 O 3 ], while the ratio K = [SiO 2 ]/[B 2 O 3 ] was kept constant and equal to 2.12. Spectra were measured at room temperature in samples with 0.43 ≤ R ≤ 1.68, and the effect of the modifier content was clearly observed in these glasses, only in partial agreement with previous literature results. In particular, the formation in the glass of sodium-danburite units Na 2 O·B 2 O 3 ·2SiO 2 was postulated. This feature led to a new assessment of R*, the critical value of R above which every new alkali atom added to the system breaks a Fo-O-Fo (Fo = glass former) bridge causing depolymerization of the glass. A revised formula is proposed to obtain the value of R* as a function of K. Raman spectra measured at high temperature yielded important information about the temperaturedependent evolution of the borosilicate system. In particular, borate and borosilicate units including tetra-coordinated boron seem to be unstable at high temperature, where the formation of metaborate chains or rings is fostered. Above 1500 °C, evaporation of borate compounds is clearly observed, stemming from the small sample size.

Considerable uncertainty has surrounded the occurrence of tetrahedrally coordinated Al and B at the T site in tourmaline. Although previously detected in several tourmaline specimens, the frequency of these substitutions in nature, as well as the extent to which they occur in the tourmaline structure, is not known. Using 11 B and 27 Al MAS NMR spectroscopy, we have investigated the presence of B and Al at the T site in 50 inclusion-free tourmaline specimens of low transition-metal content and different species (elbaite, “fluor-elbaite” , liddicoatite, dravite, uvite, olenite, and magnesiofoitite) from different localities worldwide. Chemical shifts of [4] B and [3] B in 11 B spectra, and [4] Al and [6] Al in 27 Al spectra, are well resolved, allowing detection of even small amounts of T-site constituents. In the observed spectra, [4] B and [3] B peaks are located at 0 and 18-20 ppm, respectively, with the greatest intensity corresponding to [3] B (=3 apfu). In 27 Al spectra, [4] Al and [6] Al bands are located at 68-72 and 0 ppm, respectively, with the greater intensity corresponding to [6]Al. However, inadequate separation of Y Al and Z Al precludes resolution of these two bands. Simulation of 11 B MAS NMR spectra shows that tetrahedrally and trigonally coordinated B can be readily distinguished at 14.1 T and that a [4] B content of 0.0-0.5 apfu is common in tourmaline containing low amounts of paramagnetic species. 27 Al MAS NMR spectra show that Al is also a common constituent of the T site in tourmaline. Determination of [4] Al content by peak-area integration commonly shows values of 0.0-0.5 apfu. Furthermore, the chemical shift of the 27 Al tetrahedral peak is sensitive to local order at the adjacent Y and Z octahedra, where [4] Al- Y Mg 3 and [4] Al- Y (Al,Li) 3 arrangements result in peaks located at ~65 and ~75 ppm, respectively. Both 11 B MAS NMR and 27 Al MAS NMR spectra show peak broadening as a function of transition-metal content (i.e., Mn 2+ + Fe 2+ = 0.01-0.30 apfu) in the host tourmaline. In 11 B spectra, broadening and loss of intensity of the [3] B signal ultimately obscures the signal corresponding to [4] B, increasing the limit of detection of [4] B in tourmaline. Our results clearly show that all combinations of Si, Al, and B: T = (Al, Si) 6 , T = (B, Si) 6 , T = (Al, B, Si) 6 , and T = Si 6 apfu, are common in natural tourmalines.

Lattice parameters for gold nanocrystals, quartz, and vesuvianite have been determined by electron diffraction in routine transmission electron microscopy (TEM) work, with precision and accuracy near to 0.1%, after correction for elliptical distortion. The distortion, measured in three different microscopes, is constant for each microscope and may be easily eliminated. Variable camera constants have been avoided by positioning the oriented specimen on the eucentric plane and using parallel illumination. The current flowing in the first intermediate lens was kept fixed, assuring constant conditions of the TEM projecting system, with no further diffraction focus applied. Application of this method to micas from metamorphic rocks produced deviations between measured and expected values up to 0.8%. Although easy species distinction is still possible, minor crystal chemical differences within the sample may be lost. Likely causes of these deviations are the possible heterogeneous samples, as well as beam damage leading to cation loss with subsequent variation in basal spacings.

Two monazite generations (M1; M2) were distinguished in a kyanite-garnet gneiss from the UHP terrain of the Pohorje Mountains, Slovenia. P-T estimates reveal a peak event at 760 °C/2.6 GPa and isothermal decompression down to 700 °C/0.6 GPa. M1 type provides a Th-U-Pb mean date of 100 ± 6 Ma, ThO 2 contents between 3-7 wt%, Y 2 O 3 values <0.3 wt%, and La/Nd ratios (1.2-1.4) that are clearly higher than for the whole-rock La/Nd (1.1). The absence of Y zoning in M1 and the lack of monazite inclusions in garnet indicate that M1 formed after the main stage of garnet growth (>1.2 MPa), probably close to the P-T peak. M2 type is slightly younger than M1 (74 ± 16 Ma), and has a lower La/Nd (0.3-0.9), lower ThO 2 (0.1-5 wt%), and higher Y 2 O 3 (up to 3.2 wt%). Most M2 monazites occur as tiny needles within apatite (subtype M2-a) or along apatite margins (M2-b). Parasitic growth of M2-a and -b from apatite is supported by its low ThO 2 (<1 wt%) and La/Nd (<0.5). Isolated matrix grains (M2-c) and overgrowths around M1 (M2-d) have slightly higher La/Nd (0.5-0.9) and higher ThO 2 (5 wt%) and were supplied from an apatite and M1 source. Elevated yttrium suggests that M2 formed during decompression, when garnet was consumed and Y was released. These observations imply that at UHP conditions MREE-rich apatite coexisted with low-MREE M1 monazite and reacted during decompression to Ca-F-apatite plus MREE-rich M2 monazite. This provides strong arguments that REE-partitioning between apatite and monazite is pressure-dependent.

We have explored the exchange of Cs for interlayer Na in birnessite using several techniques, including transmission electron microscopy (TEM) and time-resolved synchrotron X-ray diffraction (XRD). Our goal was to test which of two possible exchange mechanisms is operative during the reaction: (1) diffusion of cations in and out of the interlayer or (2) dissolution of Na-birnessite and reprecipitation of Cs-birnessite. The appearance of distinct XRD peaks for Na- and Cs-rich phases in partially exchanged samples offered support for a simple diffusion model, but it was inconsistent with the compositional and crystallographic homogeneity of (Na,Cs)-birnessite platelets from core to rim as ascertained by TEM. Time-resolved XRD revealed systematic changes in the structure of the emergent Cs-rich birnessite phase during exchange, in conflict with a dissolution and reprecipitation model. Instead, we propose that exchange occurred by sequential delamination of Mn oxide octahedral sheets. Exfoliation of a given interlayer region allowed for wholesale replacement of Na by Cs and was rapidly followed by reassembly. This model accounts for the rapidity of metal exchange in birnessite, the co-existence of distinct Na- and Cs-birnessite phases during the process of exchange, and the uniformly mixed Na- and Cs-compositions ascertained from point analyses by selected area electron diffraction and energy dispersive spectroscopy of partially exchanged grains.

Phosphorous-rich coffinite, U(Si,P)O 4 ·H 2 O, from the natural nuclear reactor at Bangombé, Gabon (depth 12.25 m), has been examined as an important primary mineral and alteration product of uraninite under reducing conditions. Based on electron microprobe analyses (EMPA) and textural relationships, two distinct types of coffinite have been identified: (1) P-coffinite- (i) [with P 2 O 5 and (REE+Y) 2 O 3 as high as 9.84 and 8.66 wt%, respectively] replaces uraninite and has a chemical formula of (U 0.84±0.05 Y,REE 0.12±0.02 Ca 0.10±0.02 Th 0.003±0.002 ) Σ1.07±0.06 (Si 0.41±0.07 P 0.34±0.03 S 0.08±0.05 ) Σ0.84±0.05 O 4 and; (2) coffinite-(ii), lacks uraninite inclusions [with P 2 O 5 and (Y+REE) 2 O 3 up to 1.45 and 1.79 wt%, respectively] and has a chemical formula of (U 0.78±0.02 Ca 0.05±0.003 Y,REE 0.03±0.01 Th 0.002±0.001 ) Σ0.87±0.02 (Si 1.02±0.02 P 0.06±0.01 ) Σ1.08±0.01 O 4 . The EMPA elemental maps reveal a homogeneous distribution of P, Si, Nd, and U in P-coffinite-(i). Charge-balance calculations indicate that S substitution for Si in the coffinite structure results in the enhancement of the incorporation of P and REEs. High EMPA totals (95-100 wt%), suggest that water is not an essential component of the coffinite structure. High-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) reveal the decomposition of single crystals of uraninite into 50 nm grains during the alteration process to P-coffinite-(i). High-resolution transmission electron microscopy (HRTEM) and selected area electron diffraction (SAED) patterns confirm that P-coffinite-(i) has the coffinite structure. However, P-coffinite-(i) has a large amorphous fraction probably due to radiation damage caused by α-decay events. The calculated cumulative dose for both types of coffinite varies from 1.6-1.9 × 10 18 (α-decay events/mg), which is equivalent to 134.2-161.5 displacement per atom (dpa) that have accumulated most probably during the past 800 million years.

Tistarite, ideally Ti 2 O 3 , is a new member of the corundum-hematite group. It is found as one subhedral crystal in a cluster of micrometer-sized refractory grains along with khamrabaevite (TiC), rutile, and corundum crystals within a chondrule from the Allende meteorite. The mean chemical composition determined by electron microprobe analysis is (wt%) Ti 2 O 3 94.94, MgO 2.06, Al 2 O 3 1.50, ZrO 2 0.44, FeO 0.24, CaO 0.10, Cr 2 O 3 0.06, sum 99.34. The empirical formula calculated on the basis of 3 O atoms is (Ti 3+ 1.90 Mg 0.07 Al 0.04 Zr 0.01 ) Σ2.02 O 3 . Tistarite is rhombohedral, R3̄c; a = 5.158 Å, c = 13.611 Å, V = 313.61 Å 3 , and Z = 6. Its electron back-scatter diffraction pattern matches that of synthetic Ti 2 O 3 with the R3̄c structure. The strongest calculated X-ray powder diffraction lines from the synthetic Ti 2 O3 data are [d spacing in Å (I) hkl]: 3.734 (84) (012), 2.707 (88) (104), 2.579 (90) (110), 2.242 (38) (113), 1.867 (33) (024), 1.703 (100) (116), 1.512 (28) (214), 1.489 (46) (300), 1.121 (20) (226), 0.896 (25) (416). The mineral is named after the composition “Ti” and the word “star,” implying that this new refractory mineral is among the first solids formed in the solar system.

Davisite, ideally CaScAlSiO 6 , is a new member of the Ca clinopyroxene group, where Sc 3+ is dominant in the M1 site. It occurs as micro-sized crystals along with perovskite and spinel in an ultra-refractory inclusion from the Allende meteorite. The mean chemical composition determined by electron microprobe analysis is (wt%) SiO 2 26.24, CaO 23.55, Al 2 O 3 21.05, Sc 2 O 3 14.70, TiO 2 (total) 8.66, MgO 2.82, ZrO 2 2.00, Y 2 O 3 0.56, V 2 O 3 0.55, FeO 0.30, Dy 2 O 3 0.27, Gd 2 O 3 0.13, Er 2 O 3 0.08, sum 100.91. Its empirical formula calculated on the basis of 6 O atoms is Ca 0.99 (Sc 0.50 Ti 3+ 0.16 Mg 0.16 Ti 4+ 0.10 Zr 0.04 V 3+ 0.02 Fe 2+ 0.01 Y 0.01 ) Σ1.00 (Si 1.03 Al 0.97 ) Σ2.00 O 6 . Davisite is monoclinic, C2/c; a = 9.884 Å, b = 8.988 Å, c = 5.446 Å, β =105.86°, V = 465.39 Å 3 , and Z = 4. Its electron back-scattered diffraction pattern is an excellent match to that of synthetic CaScAlSiO 6 with the C2/c structure. The strongest calculated X-ray powder diffraction lines are [d spacing in Å (I) (hkl)]: 3.039 (100) (2̄21), 2.989 (31) (310), 2.943 (18) (3̄11), 2.619 (40) (002), 2.600 (26) (1̄31), 2.564 (47) (221), 2.159 (18) (3̄31), 2.137 (15) (4̄21), 1.676 (20) (2̄23), and 1.444 (18) (531). The name is for Andrew M. Davis, a cosmochemist at the University of Chicago, Illinois.

The position of the infrared absorption band near 4500 cm -1 shifts from 4520 to 4440 cm -1 as the Al content increases along the SiO 2 -NaAlSiO 2 join and closely follows the sum of the positions of the Raman bands near 900 and 3600 cm -1 . This confirms the idea that the 4500 cm -1 band is a combination band of the fundamental (Si,Al)-OH vibration near 900 cm -1 and the fundamental O-H stretching vibration near 3600 cm -1 . As a consequence, the 4500 cm -1 band should not be used to quantify the water speciation for glass compositions for which significant amounts of free hydroxyls are expected as these do not contribute to the band’s intensity.