Volume 85, Issue 9

Studies of crystal size distributions (CSD) can reveal much about how rocks solidify and under what conditions. Data from two-dimensional sections can be readily acquired at many different scales, from electron microscope images, thin sections, slabs, outcrops, and so on, but the conversion to true, three-dimensional values is complex. The widely used Wager method does not have a good theoretical basis and does not give accurate results. A modification of the Saltikov correction method is proposed here that is more accurate and can account for different crystal shapes and fabrics. Population densities determined by this method differ by factors of 0.02 to 100 from those determined by the Wager method. Published CSDs determined using other methods can be recalculated if the crystal shape and fabric parameters can be estimated. The method has been incorporated into a new program, CSDCorrections.

The solubilities of Ar and Kr in supercooled melts in the system SiO 2 -NaAlSi 2 O 6 have been studied at temperatures of 1200 °C (SiO 2 ), 1000 and 750 °C (NaAlSi 3 O 8 ), and 800 °C (NaAlSi 2 O 6 ) at pressures from 200 to 6000 bar. Gas sorption experiments were performed in an internally heated pressure vessel, and noble gas concentrations in quenched samples were analyzed by absolute analytical techniques [gas chromatography (GC), and thermogravimetry (TG), gravimetry]. Comparison of the results from this study with other absolute techniques (Rutherford-backscattering, mass spectrometry) is possible for silica and exhibit excellent agreement. This agreement leads us to conclude that our silica glasses with an Ar concentration of 1.02 wt% and a Kr concentration of 0.54 wt% are suitable as standard materials. Solubility increases linearly with increasing pressure for all three compositions. In SiO 2 the solubility of Ar and Kr increases with pressure with nearly parallel slopes, whereas Ar and Kr exhibit diverging slopes in NaAlSi 3 O 8 and NaAlSi 2 O 6 . The linear relationship between fugacity and dissolved gas below 2-3 kbar, indicates that Henry’s law constants (in units of 106 bar) for Ar in SiO 2 , NaAlSi 3 O 8 , and NaAlSi 2 O 6 are 0.94, 3.15, and 5.71, respectively, and for Kr are 1.38, 5.69, and 9.28, respectively. The experimental results can be modeled by two alternative thermodynamic approaches. First, assuming mixing of noble gases with the network of the melt, partial molar volumes (in cm 3 /mol) can be calculated in SiO 2 , NaAlSi 3 O 8 , and NaAlSi 2 O 6 for Ar as 25.9 ± 1.3, 21.1 ± 1.6, and 21.1 ± 0.7, respectively and for Kr as 24.3 ± 1.8, 27.6 ± 0.8, and 24.7 ± 0.9, respectively. Second, assuming the volume change upon solution to be zero, a Langmuir- Isotherm can be applied yielding saturation levels M (in units of 1020 sites/cm 3 melt) in SiO 2 , NaAlSi 3 O 8 , and NaAlSi 2 O 6 for Ar as 3.82 ± 0.6, 1.33 ± 0.2, and 0.72 ± 0.08, respectively and for Kr as 3.49 ± 0.7, 0.56 ± 0.05, and 0.42 ± 0.06, respectively. Highest solubility is observed in all three melts for Ar. Solubility decreases with increasing substitution of Si 4+ by Na + + Al 3+ . The dependence of solubility on composition supports the structural model for melts along the join SiO 2 -NaAlSi 2 O 6 based on a stuffed tridymite-like structure.

The solubility of H 2 O in three melt compositions along the haploandesite join Na 2 Si 4 O 9 - Na 2 (NaAl) 4 O 9 (0, 3, and 6 mol% Al 2 O 3 ) was determined as a function of pressure and temperature from 0.8 to 2.0 GPa and 1000 to 1300 °C. Water solubility is a linear (or near-linear) positive function of pressure (16-18 mol% H 2 O/GPa) at constant temperature, and a negative near-linear function of temperature (1-2 mol% H 2 O/100 °C) at constant pressure. The solubility is correlated negatively with Al 2 O 3 content of the melts. Partial molar volume of H 2 O in the melt, V̅ H2O melt , was derived from solubility isotherms (1000, 1100, 1200, and 1300 °C) at 0.8, 1.05, 1.3, 1.65, and 2.0 GPa pressure. Values range between 7.8 and 12.8 cm 3 /mol, and decrease with increasing Al 2 O 3 content. In the pressure-temperature range studied, (∂V̅ H₂O melt /∂T) p ranges from -7.1 ± 0.810 -3 to -5.6 ± 1.3·10 -3 cm 3 /mol °C, becoming slightly less negative as the melts become more aluminous. The V̅ H₂O melt values were combined with published partial molar volume information for anhydrous oxides in silicate melts to estimate densities of water-rich dacitic magmas in shallow magma chambers associated with explosive volcanism. For a chamber of constant bulk composition during a comparatively short explosive event, such as that of Mount Pinatubo in June 1991 or Mount St. Helens in May 1980, the average density of the magma after eruption is ~3% higher than before the eruption occurred. Furthermore, because of removal of overburden during an eruption, the H 2 O saturation value of remaining magma is less than that prior to eruption. From density calculations of the residual hydrous magma after eruption, its density decreases from top to bottom in the magma chamber. Consequently, this magma is gravitationally unstable.

Atomistic simulation techniques are used to investigate the effect of proton-containing defects on the structure and stabilities of a range of grain boundaries of forsterite. We study two series of stepped {010} tilt boundaries that are at 90° to each other: one with the {100} plane as step wall and the other with the {001} plane. Each series consist of several grain boundaries with increasing terrace area (i.e., decreasing boundary angle). The ratios of boundary and surface energies γ b /γ s and γ b /γ {010} with boundary angle show maxima at a boundary angle ϕ = ~30° and minima at ϕ = 0° and ~60°. The adhesion energies of the two series show a minimum at low boundary angle (ϕ = 20-30°), indicating that there is an optimum size for the {010} terrace area, where the relative stabilities of grain boundary and related surface make separation of the boundary into the free surfaces energetically least expensive. Dissociative adsorption of water molecules in the bulk crystal is preferred at the M2 site, but is an endothermic process with a calculated hydration energy of +119 kJ/mol. Hydration of the grain boundaries on the other hand is energetically favorable with hydration energies tending toward -80 kJ/mol compared with a surface hydration energy at the planar {010} surface of -90 kJ/mol. We also investigated the adsorption of protons at cation defects, by modeling the process of replacement of Mg ions in the bulk and along the grain boundaries by two H + ions each. Replacement of Mg 2+ in the bulk crystal occurs preferentially at the M1 site and is calculated to be exothermic with a replacement energy of -78 kJ/mol; Mg 2+ replacement along both series of grain boundaries is also an exothermic process. The replacement energies tend to a constant value of approximately -210 kJ/mol with increasing terrace area. (cf., the calculated value for the planar {010} surface of -230.4 kJ/mol).

An atomistic computer simulation study has been undertaken to determine the energies associated with different mechanisms of Fe 3+ incorporation into forsterite and wadsleyite, and to determine why Fe 3+ substitutions into some mantle minerals (such as forsterite) are less favorable than in others. We have also compared our results to previous results obtained for perovskite. In all three phases, the most favorable substitution mechanism is Fe 3+ entering both the Mg and Si sites. In forsterite, the energy of incorporation is 2 eV less favorable than the same mechanism in perovskite and 1 eV less favorable than in wadsleyite. These differences are due to significantly different energies for the substitution of Fe 3+ into the Si site in each of these silicates. This substitution was most favorable in perovskite (octahedral site), less favorable in wadsleyite (tetrahedral site) and least favorable in forsterite (also tetrahedral). The energy difference between forsterite and wadsleyite was found to be the result of structural effects. The linked tetrahedra in wadsleyite are able to distort more easily to accommodate Fe 3+ than the isolated SiO 4 tetrahedra in forsterite.

Molecular dynamics (MD) simulations of the thermodynamic properties of Al/Si-ordered plagioclase were performed for two different crystallographic configurations: a solid solution structure and a lamellar structure. MD-simulated plagioclase feldspar solid solution has three space groups -C1̅ - , I1̅-, and P1̅ - at 300 K, and two space groups -C1̅ - and I1̅ - at higher temperatures. In addition, the MD-simulated, composition-dependent I1̅-C1̅ and P1̅-I1̅ phase transitions are non-first order. However, the temperature-dependent I1̅ - C1̅ phase transition is not observed for any plagioclase composition. The calculated excess enthalpy and excess Gibbs free energy of the MD-simulated solid solution structure are convex upward, and for a given plagioclase composition the excess Gibbs free energy of the MD-simulated lamellar structure is smaller than that of the MD-simulated solid solution structure. These results indicate that a lamellar structure (created by phase separation) consisting of fully ordered albite and anorthite is more stable than the solid solution structure, which consists of randomly distributed clusters of both phases. Therefore, when plagioclase feldspar is heated for a long period of time at a low temperature, phase separation between pure albite and pure anorthite can be expected if equilibrium is achieved.

The equilibrium distribution of Mg and Al between the tetrahedral and octahedral sites of a flux grown, stoichiometric MgAl 2 O 4 spinel was investigated between 600 and 1100 °C by single-crystal X-ray diffraction of quenched samples. The cation distribution for both ordering and disordering runs was obtained by minimizing accurate crystallographic parameters and effective ionic radii. Along with the variation of the degree of inversion from 0.18 to 0.29 between 600 and 1100 °C, both unit cell and oxygen positional parameter decreased linearly. Multiple non-linear least-squares fit of our data with the thermodynamic model of O’Neill and Navrotsky (1983) gave α = 23 ± 2 KJ/mol and β = 13 ± 4 KJ/mol. The influence of both cation inversion and thermal expansion on T-O and MO bond length variation was determined by comparison of our data with previous in situ studies. In the thermal interval investigated, the inversion accounts for change of -0.014 Å for T-O and +0.007 Å for M-O. Mean linear polyhedral thermal expansion coefficients of 6.5 × 10-6 °C-1 and 8.9 × 10 -6 °C -1 were calculated for T and M sites, respectively.

Al K-edge XANES spectra were recorded for six tektites to obtain information on the local structure around Al. Albite, grossular, and andalusite were used as reference materials for Al in fourfold, sixfold, and five+sixfold coordination, respectively. The spectra of the tektites are very similar, except for the intensity of the main absorption edge. Comparison of the spectral features of the tektite glasses with those of the reference materials indicates that Al is tetrahedrally coordinated.

A Rietveld refinement of Cr-doped mullite, Cr 0.5 Al 3.92 Si 1.58 O 9.79 , orthorhombic, space group Pbam [a = 7.56712(6) Å, b = 7.70909(6) Å, c = 2.90211(2) Å, V = 169.30 Å 3 ] revealed Cr to reside predominantly in the octahedrally coordinated M1 site. The mean M1-O distance of 1.935 Å is that expected from Al 3+ and Cr 3+ molar fractions obtained from the chemical analysis and structure refinement. The small displacement factors of the T* and Oc* sites indicate deficiencies in the electron densities which could be compensated by assuming additional Cr atoms. The predominant Cr incorporation into the M1 site causes an expansion of the octahedral bonds which is directly related to the observed lengthening of the c edge. The strong expansion of the long and elastic octahedral M1-Od bond in Cr-doped mullite, which would affect the a and b edges, is partly absorbed by a shortening of tetrahedral bonds.

Comparison of experimental single-crystal electron diffraction patterns of synthetic two-line ferrihydrite (2LFh) with simulated single-crystal electron-diffraction patterns indicates that a synthetic 2LFh sample contains highly disordered material and nanocrystals with structures based on hexagonal (ABAB) and cubic (ABC) stacking of close-packed layers of O 2 - and OH- ions. An apparently continuous variation in ordering exists between the highly disordered material and each of the crystalline structures, suggesting that both nanocrystalline structures represent local extremes of three-dimensional ordering. Experimental diffraction patterns were obtained using electron nanodiffraction, a technique in which the finely focused beam from a field-emission gun in an electron microscope can be used to produce diffraction patterns from areas <1 nm across. Nanodiffraction patterns from the highly disordered material have diffuse streaks rather than distinct reflections, and are consistent with a two-dimensional structure that consists of close-packed anionic layers with essentially complete stacking disorder and nearly random distribution of Fe atoms. The structure with cubic stacking is similar to maghemite and has ~25% of the Fe in tetrahedral sites. The structure with hexagonal stacking consists of double chains of face-sharing Fe octahedra; each octahedron shares one face, two edges, and three corners with adjacent octahedra. Previous results from transmission electron microscopy, powder X-ray and electron diffraction, and synchrotron-based techniques reflect the overall high degree of structural disorder rather than the characteristics of the maghemite-like and double-chain structures

A transmission electron microscopy (TEM) study of gaudefroyite, Ca 8 Mn 3+ 6 [(BO 3 ) 6 (CO 3 ) 2 O 6 ], from the Wessels mine, Kalahari manganese field, South Africa, shows strong, continuous streaking in selected-area electron diffraction (SAED) patterns. The pseudo-hexagonal parameters for gaudefroyite are a = 10.606(1) and c = 5.896(1) Å. The symmetry of gaudefroyite is lower than hexagonal because of the streaked reflections. The CO 3 groups in the large 6 3 channels are ordered in two different ways and this ordering gives rise to two different types of channels, A and B. The A and B channels occur in a 1:1 ratio and these channels are partially ordered and give rise to the continuous streak reflections in gaudefroyite.

Samples of phengite-3T, (K 0.93 Na 0.01 )(Al 1.44 Mg 0.56 Ti 0.02 )(Si 3.54 Al 0.46 )O 10 (OH 1.93 F 0.07 ), which formed at about 3 GPa and 1000 K in ultrahigh-pressure metamorphic rocks of the Dora-Maira Massif (Western Alps, Italy), were investigated by HRTEM, AEM, EMPA, and SEM. The matrix of phengite- 3T, which is almost free of stacking faults, contains single-crystal α-quartz platelets 100-700 Å thick that are confined by (001) planes of phengite. In the vicinity of quartz, the 30 Å period of the phengite-3T matrix is faulted by short sequences with about 19 Å periodicity, interpreted as talc-2M. The occurrence of these phases is not connected with any defects in the host phengite nor is their spatial distribution homogeneous on the TEM scale. The shape control exerted by phengite on the quartz crystals, the absence of deformation around them, and the nearby presence of an interlayered 19 Å phase, suggest that quartz and talc may have been produced within mica by a reaction of the type: 3 alumino-celadonite + 2 H + = 1 muscovite + 1 talc + 5 quartz + 2 H 2 O + 2 K + which leads to a less celadonite-rich phengite during decompression, after the rock had left the coesite stability field. In addition, examination by optical microscopy along [001] of “thick” sections of the Dora-Maira phengite and of phengite samples from other high-pressure terranes (Monte Mucrone and Sesia Zone in the Italian Western Alps, Dabie Mountains in central China), revealed the presence of micrometer-wide, amoeboid platelets of quartz interlayered at various depths parallel to (001) of micas. In spite of the different observation scale, these are interpreted as the same reaction product as identified by HRTEM. These new observations show that high-pressure white micas may not be homogeneous and should be examined more carefully. Some consequences for thermobarometry of such heterogeneity and intracrystalline re-organization during decompression are considered; they depend on the resolution of the analytical method employed. Implications for thermochronometry still have to be evaluated.

Near-atomic-resolution transmission electron microscopy was used to investigate the chloritization mechanisms of biotite in a granitic rock and to reveal polytypic details of the resulting chlorite. Comparison of stacking sequences in 2M 1 and longer period biotite polytypes with sequences in areas containing chlorite layers revealed that typically, two biotite layers transform to one chlorite layer, losing two potassium interlayer sheets and two tetrahedral sheets. In some cases, more than two biotite layers are replaced by one chlorite layer. Less commonly, a potassium interlayer sheet is replaced by a brucite-like sheet. One biotite layer is transformed to one serpentine layer via loss of a potassium interlayer and a tetrahedral sheet in places. Based on the relative frequency of the two chlorite layers to one biotite layer vs. one biotite layer to one chlorite layer mechanisms, the net result of chloritization is a considerable volume decrease along c*. Near-atomic-resolution images recorded down [010], [310], or [31̄0] revealed that the chlorite polytype in the biotite-chlorite interstratifications is predominantly IIbb. However mixtures of Ibb, Iab, IIab, and IIbb also occur. The “aa” stacking sequences were never found. Chlorite polytypes may be determined in part by the chloritization mechanisms and in part by relief of local shear by a/3 displacements.

The dissolution of individual hectorite (a trioctahedral smectite) particles has been observed at molecular scales in acidic aqueous solution with atomic force microscopy (AFM). A new sample preparation technique was used to attach nanometer-sized hectorite particles to a mica substrate. The reactive surface area of individual hectorite particles was identified and its change during a dissolution experiment was quantified. The dissolution of hectorite under pH 2 conditions takes place exclusively at the edge surfaces. In contrast, the basal surface is completely unreactive within the investigated time scale of several hours. The short edges of the hectorite laths were found to react somewhat more quickly than the long edges. The edge surface area represents 1.5-3.3% of the total surface area. The total surface area has been determined from the actual particle dimensions derived from AFM data to be 730 m 2 /g. The dissolution rate normalized to the reactive edge surface area (ESA) has been determined to be 7.3 × 10 -9 mol hectorite/(m 2 -s), which represents a total surface area (TSA) normalized dissolution rate of 1.9 × 10 -10 mol hectorite/(m 2 -s). The ESA/TSA ratio increases by about 15% within 1 h exposure to a pH 2 aqueous solution at 22 °C.

Minor amounts of clay minerals precipitated from aqueous solution can be rapidly identified and quantified in a mineral mixture with powder X-ray diffraction using a rotating-anode source and a position-sensitive detector. For the case of gibbsite precipitated on a kaolinite powder substrate we demonstrate a simple method having a minimum detection limit of 0.1 wt%, using pure gibbsite as the intensity reference in mechanical mixtures of gibbsite and kaolinite. The amount of gibbsite precipitated onto kaolinite at 80 °C, pH 3 is higher when determined from solution chemistry than from the X-ray method, and the difference in amounts increases with increasing Al concentration in solution. This discrepancy can be explained by assuming that a fraction of the precipitated material is effectively invisible to the X-ray diffraction technique, either due to a small diffracting domain size along the gibbsite [001] direction or formation of an Al-phase that is amorphous to X-rays. This method should be generally useful for a range of mineral mixtures where at least one intense reflection for the phase of interest is not obscured. The ability to identify, characterize, and quantify trace phases by X-ray diffraction, especially when combined with surface analysis by electron or atomic force imaging, is an important complement to the conventional approach of monitoring solution composition in growth kinetics experiments.

Illite-smectite (I-S) from Cambrian black shale of both early diagenetic and anchimetamorphic grade was investigated to determine the mechanism of the clay transformation. The layer sequences, the distribution of thicknesses of coherent scattering domains (CSDs), and the three-dimensional ordering were determined by X-ray diffraction (XRD). The proportions of cis-vacant (cv) and transvacant (tv) 2:1 layers were determined by thermal analysis and the proportion and distribution of interlayer ammonium by XRD and by infared spectroscopy (IR). The structural formulae were determined from total chemical analysis, and Mössbauer and 27 Al NMR spectroscopies, and the particle shape and size investigated by atomic force microscopy (AFM). In the early diagenetic samples, the I-S is composed of two phases, one of which contains 0.05 and the other 0.25 smectite (S) interlayers. The first phase does not change during metamorphism. In the second phase, 0.20 S are converted to tobelite (T) layers through fixation of NH 4 + , but the I layers are not changed. Simultaneously, the proportion of cv layers changes from 0.18 to 0.02, and the tetrahedral substitution of Al for Si is parallel to the increase in T layers. All I interlayers contain 0.75K per O 10 (OH) 2 . Furthermore, the metamorphism results in increasing mean thickness of CSDs from 5.1-6.8 nm for the lowdiagenetic samples to 6.7-8.4 nm for the anchimetamorphic samples. We conclude that the tobelitization was accompanied by transformation of cv to tv 2:1 layers adjacent to the smectite interlayers, and formation of tv layers adjacent to the newformed tobelite interlayers in otherwise intact crystallites. This mechanism only partly resembles the tobelitization previously observed in the Upper Jurassic North Sea oil source rocks. I-S in these rocks contained tv 2:1 layers and T interlayers formed through solid-state Al for Si substitution in the tetrahedral sheet and by ammonium fixation in the corresponding interlayers. These different mechanisms are probably because the North Sea I-S originated from weathered illite, like the Cambrian high-illitic phase, whereas the Cambrian low-illitic phase undergoing the transformation originated from cv smectite of volcanic origin. The results indicate that the illitization in oil source rocks is linked to oil generation, and that it deviates from the illitization in other rocks because of the supply of ammonium formed during oil generation and the fixation of this ammonium in the former smectite interlayers.

Unit-cell parameters, optical properties, and chemical compositions have been measured for 103 samples in the tremolite-actinolite-ferro-actinolite series. The average values of the non-essential constituents are: T Al = 0.10(11), C Al = 0.06(6), B (Fe, Mn, Mg) = 0.09(7), B Na = 0.04(5), A Na = 0.09(9), and Cr, Ti, and K ≅ 0. Asbestiform actinolite samples have lower Al contents than massive or “byssolitic” actinolite samples. Unit-cell parameters for end members tremolite and ferro-actinolite based on regressions of the data are: a = 9.841 ± 0.003 Å, 10.021 ± 0.011 Å; b = 18.055 ± 0.004 Å, 18.353 ± 0.018 Å; c = 5.278 ± 0.001 Å, 5.315 ± 0.003 Å; and cell volume = 906.6 ± 0.5 Å 3 , 944 ± 2 Å 3 . The changes in a, b, and cell volume with ferro-actinolite substitution are modeled with quadratic functions, and the change in c with ferro-actinolite substitution is modeled with a linear function. There is a positive correlation between c and Al that results in a discrimination between asbestiform and massive or “byssolitic” habits for c and for the refractive indices. The principal refractive indices n γ and n β are linear with respect to ferro-actinolite substitution, but n α is best modeled by two lines with a change in slope between 26 and 32% ferro-actinolite. Birefringence and extinction angle also change between 26 and 32% ferro-actinolite. The predicted end-member values of the principal refractive indices for tremolite and ferro-actinolite are: n α = 1.602 ± 0.001, 1.661 ± 0.005; n β = 1.621 ± 0.001, 1.692 ± 0.004; and n γ = 1.631 ± 0.001, 1.700 ± 0.003. There is a discontinuity in a at approximately 11% ferro-actinolite that is accompanied by a drop in Ca. There are also indications of discontinuities in optical properties and c between 26 and 32% ferro-actinolite that may be due to an increase in tschermakite substitution. Both discontinuities are accompanied by a decrease in the relative frequency of natural samples.

The incorporation of Fe 2+ was investigated in four natural cordierite samples. 57 Fe Mössbauer, single-crystal UV-VIS optical absorption, and X-ray absorption spectroscopies, as well as X-ray single-crystal diffraction were used. Mössbauer, optical, and XAS spectroscopy show that Fe 2+ is incorporated on two different structural sites in two Mg-rich samples. Mössbauer measurements give the best quantitative measure of the amounts of Fe 2+ , but the optical spectra are the most sensitive for determinations at low concentrations and at high-bulk Fe 2+ concentrations in cordierite. The spectroscopic data are most consistent with small amounts of Fe 2+ (i.e., 0.02 of Fe 2+ per formula unit) being located on a tetrahedral site rather than in the center (or off center) of the six-membered tetrahedral rings or in channel cavities. X-ray single-crystal refinements on two Mg-rich cordierites show a very small excess electron density on T 1 1 and not in the channels. A third refinement on a slightly more iron-rich sample shows, in contrast, no excess electron density on T 1 1. We interpret these data as indicating that small amounts of Fe 2+ (0.01 to 0.02 atoms per formula unit) replace tetrahedral Al 1 1 in cordierite, where charge balance is achieved by placing Na in the center of the six-membered rings. This substitution is consistent with the known chemistry of natural cordierites and with simple structural energetics. The identification and assignment of small amounts of Fe 2+ on T 1 1 requires spectroscopic determination or careful X-ray single-crystal refinements and cannot be achieved from composition data and structural formula calculations.

Polarized single-crystal Raman spectra at room temperature and 5 K and polarized infrared spectra at room temperature were obtained from four natural cordierites of different compositions in the wave number region of the CO 2 symmetric stretching vibration and the H 2 O stretching vibrations. The CO 2 molecules are preferentially aligned parallel to the X axis, consistent with results from X-ray diffraction and optical studies. The CO 2 contents of six natural cordierites, previously studied by powder IR methods (Vry et al. 1990), were determined via Raman spectroscopy. A linear relationship was found between CO 2 content and the Raman intensity ratio of the normalized CO 2 stretching mode against a Si-O stretching mode. This permits a determination of the CO 2 contents in cordierite using micro-Raman measurements. The internal stretching modes between 3500 and 3800 cm -1 were assigned to various types of H 2 O molecules occurring in the channel cavity. Three different orientations of H 2 O molecules that have no interactions with alkali cations located at 0,0,0 in the sixmembered tetrahedral rings are classified in a static model as Class I H 2 O molecules. The H···H vector for two of them is parallel to [001], and their molecular planes lie in the XZ and YZ crystal planes. The third type has its H···H vector directed along the X axis and its molecular plane lies in the XZ plane. Two other types of H 2 O have interactions with the alkali cations located at 0,0,0. They are classified as Class II H 2 O. They distinguish themselves by the number of H 2 O molecules bonded to the alkali atoms. The formation of weak hydrogen bonds at low temperatures may explain the appearance of some Raman stretching modes below 200 K. The H 2 O molecules of Class I-Type I/II are probably dynamically disordered about [001] hopping between orientations in the XY and XZ planes down to 5 K. Class II H 2 O may also be disordered, but more measurements are required to describe its dynamic behavior.

The crystal structure and M-site populations of a series of micas-1M from miarolitic pegmatites that formed within host granitic rocks of the Precambrian, anorogenic Pikes Peak batholith, central Colorado, were determined by single-crystal X-ray diffraction data. Crystals fall in the polylithionitesiderophyllite- annite field, being 0 ≤ Li ≤ 2.82, 0.90 ≤ Fe total ≤ 5.00, 0.26 ≤ [6] Al ≤ 2.23 apfu. Ordering of trivalent cations (mainly Al 3+ ) is revealed in a cis-octahedral site (M2 or M3), which leads to a lowering of the layer symmetry from C12/m(1) (siderophyllite and annite crystals) to C12(1) diperiodic group (lithian siderophyllite and ferroan polylithionite crystals). On the basis of mean bond length, the ordering scheme of octahedral cations is mostly meso-octahedral, whereas the mean electron count at each M site suggests both meso- and hetero-octahedral ordering, the calculated mean atomic numbers being M1 = M3 ≠ M2, M2 = M3 ≠ M1 and M1 ≠ M2 ≠ M3. As the siderophyllite content increases, so do the a, b, and c unit-cell parameters, as well as the refractive indices, primarily n β . The tetrahedral rotation angle, α, is generally small (1.51 ≤ α ≤ 5.04°) and roughly increases with polylithionite content, whereas the basal oxygen out-of-plane tilting, Δz, is sensitive both to octahedral composition and degree of order (0.0 ≤ Δz ≤ 0.009 Å for siderophyllite and annite, 0.058 ≤ Δz ≤ 0.144 Å for lithian siderophyllite and ferroan polylithionite crystals).

Thallium aluminosilicate, TlAlSiO 4 , synthesized hydrothermally is monoclinic with space group P2 1 /n [a = 5.4095(3), b = 9.4232(7), c = 8.2629(6) Å, γ = 90.01(2) ° , V = 421.20(6) Å 3 , Z = 4]. The crystal structure was refined to an R index of 3.8% based on 1852 observed unique reflections. The compound is a unique framework silicate with a topology similar to that of the tridymite structure. The TlO 8 polyhedron resembles a truncated rectangular pyramid, and shares its edges with three adjacent AlO 4 tetrahedra, three SiO 4 tetrahedra, and six TlO 8 polyhedra. Local understaturation at the Tl position suggested by bond-valence analysis implies that lone-pair electrons are present. The geometrical data indicate that the inert pair causes distortion of the Tl-polyhedron. Polyhedral distortion analysis using the software IVTON places the lone-pair parallel to [010], pointing to the largest base of Tl polyhedron. The rule in the valence shell electron pair repulsion model that a nonbonding pair occupies more space on the “surface” of the central atom than a bonding pair supports the orientation of inert-pair electrons in thallium provided by IVTON. The remarkable structure distortion caused by the inert-pair effect explains the rarity of Tl as a major element in silicate minerals because these cannot accommodate extremely distorted polyhedra. In contrast, about forty species of Tl-sulfide minerals exist because these structures are more flexible. Furthermore this effect probably explains why atoms such as Ge 2+ , Pb 2+ , Sn 2+ , Sb 3+ , and Bi 3+ , crystallize not as silicate phases but mainly as sulfide ones in nature.

The structure of agrinierite, K 2 (Ca 0.65 Sr 0.35 )[(UO 2 ) 3 O 3 (OH) 2 ] 2 ·5H 2 O, orthorhombic, F2mm, Z = 16, a = 14.094(2), b = 14.127(2), c = 24.106(4) Å, V = 4799.6(1) Å 3 , was solved by direct methods and refined by full-matrix least-squares techniques to an agreement factor (R) of 6.55% and a goodness- of-fit (S) of 0.851 using 2710 independent observed reflections collected with MoKα X-radiation and a CCD-based detector. This layered material contains four unique U 6+ positions, each of which is part of a nearly linear (UO 2 ) 2+ uranyl ion. The U 6+ cations are further coordinated by five anions occupying the equatorial vertices of pentagonal bipyramids that are capped by the uranyl ion O atoms. The uranyl polyhedra are linked by the sharing of equatorial vertices and edges in a fashion topologically identical to the α-U 3 O 8 sheet found in billietite, protasite, becquerelite, richetite, compreignacite and masuyite. The arrangement of hydroxyl anions within the sheets varies in these minerals; that of agrinierite is identical to protasite. The cations (Ca, Sr, and K) and H 2 O reside in the interlayer region of the structure. The inclusion of Sr in the structure of agrinierite suggests that the release of radioactive 90 Sr may be impacted by incorporation into this phase if it forms in a geological repository for nuclear waste.

The crystal structure of namibite, Cu(BiO) 2 VO 4 (OH) [triclinic, space group P1̄, a = 6.210(1), b = 7.398(1), c = 7.471(1) Å, α = 90.10(1), β = 108.73(1), γ = 107.47(1)°, V = 308.22(8) Å 3 , Z = 2] was determined using single-crystal X-ray diffraction data. The refinement, based on 2140 unique reflections with F o > 4σ(F o ), gave R1 = 3.61%. The structure determination showed namibite to be triclinic-pseudomonoclinic; the previously reported C-centered monoclinic cell is a pseudocell. The structure contains two unique Bi atoms, with nine Bi-O bonds between 2.17 and 3.39 Å. The Bi coordinations show some stereochemical influence of the lone pair of electrons on Bi 3+ . A threedimensional network is formed by linkages of Bi-O polyhedra to one slightly distorted V 5+ O 4 tetrahedron that decorates chains of corner-sharing CuO 6 polyhedra extending parallel to the b axis. The CuO 6 polyhedra show the Jahn-Teller-distorted [4+2]-coordination of Cu 2+ . Layers of Bi atoms parallel to (100) alternate with layers of parallel heteropolyhedral [Cu(VO 4 )O 2 (OH)] chains. According to bond-valence calculations and geometrical considerations, the H atom of the OH group is probably involved in disordered or trifurcated H bonding. The structure of namibite represents a slightly distorted ordered variant of the monoclinic structure of brendelite, (Bi,Pb) 2 (Fe 3+ ,Fe 2+ )O 2 (OH)(PO 4 ). Further relations to synthetic Cu 3 Bi 4 V 2 O 14 , to the 7 Å chain-structure phosphates and sulfates, and to linarite are outlined.

The crystal structure of pararobertsite, Ca 2 (H 2 O) 2 [Mn 3+ 3 O 2 (PO 4 ) 3 ]·H 2 O, P2 1 /c, a = 8.814(4), b = 13.233(5), c = 11.056(4) Å, β = 101.184(7)°, V = 1265.0(9) Å 3 , Z = 4 has been solved by direct methods and refined to R = 0.042 for 1319 F o > 4σ (F o ) using MoKα X-ray data. In the structure of pararobertsite, Z-shaped edge-sharing chains of Mn3+O6 octahedra link to one another via shared O vertices and PO 4 tetrahedra to form a compact sheet of composition 2 ∞ [Mn 3+ 3 O 2 (PO 4 ) 3 ] 4- oriented parallel to {100}. The space between the compact sheets is filled with a thick open assemblage of CaO 5 (H 2 O) 2 polyhedra and isolated water molecules. The structure bears strong similarities to the mitridatite (Fe 3+ ) structure and, therefore, to its robertsite (Mn 3+ ) isotype.

Cabalzarite, M1 Ca M2 (Mg,Al,Fe 3+ ) 2 ( X AsO 4 ) 2 (H 2 O,OH) 2 , is a new mineral of the tsumcorite group occurring in altered Mn ore at the abandoned Falotta mine (Swiss Alps). Together with other arsenates, cabalzarite documents the mobility of As during the retrograde stage of the Tertiary Alpine metamorphism under lowest to sub-greenschist facies conditions. Cabalzarite crystals vary in morphology from hatchet-like to fibrous and tabular. The color is light-brownish to salmon pink or orange brown, and the average refractive index is around 1.7. Cabalzarite is chemically inhomogeneous, with the main variations occurring on the octahedral M2 site occupied by Mg, Al, Fe 3+ , and Mn 3+ . Al and Mg are the dominant M2 cations, with Al/(Al + Mg) ratios varying between 0.31 and 0.59 (92 analyses). Cabalzarite is the first member of the tsumcorite group with octahedral Mg and Al as major constituents. Single-crystal X-ray structure refinements were performed on two different crystals. Cabalzarite is monoclinic, space group C2/m, Z = 2. The cell parameters for the type crystal CABA11.5, of composition (Ca 1.00 Sr 0.02 )(Al 0.80 Mg 0.77 Fe 0.23 Mn 0.03 ) Σ1.83 (AsO 4 ) 2 (H 2 O 1.26 OH 0.74 ) 2 , are a = 8.925(2) Å, b = 6.143(1) Å, c = 7.352(1) Å, β = 115.25(3)˚, ρ calc = 3.73 g/cm 3 . The structure of cabalzarite is isomorphic with that of tsumcorite. The site M1 (Ca) is eightfold-coordinated (6 + 2) with average M1-O = 2.549 Å; M2 is octahedral with an average M2-O = 2.010 Å; and the average As-O distance of the arsenate group is 1.689 Å. Charge balance for the simultaneous occupation of M2 by two- and three-valent cations is achieved by H 2 O as well as OH contributing to the M2 coordination.

Tegengrenite, forming octahedra up to 150 μm in size, occurs associated with hausmannite, calcite, brucite, dolomite, clinohumite, kinoshitalite, native copper, barytocalcite, bindheimite, and cerussite in Mn ore from Jakobsberg, Filipstad district, Värmland, Sweden. The mineral is deep red and translucent, with a sub-adamantine luster. In reflected light it is gray and nearly isotropic, with measured reflectance values (R%) 10.4 (λ = 470 nm), 10.0 (546 nm), 9.9 (589 nm), and 9.8 (650 nm). The refractive index at λ = 589 nm is 1.92(2). There is no cleavage; fracture is conchoidal. D calc is 4.58(1) g/cm 3 . Electron-microprobe analyses gave (average of 35 points; in wt%) MgO 21.83, Al 2 O 3 0.76, SiO 2 1.70, Sb 2 O 5 36.13, TiO 2 1.40, Fe 2 O 3 0.78, MnO 25.74, Mn 2 O 3 (calculated from stoichiometry) 8.14, ZnO 2.66, sum 99.14, yielding the empirical formula Mg 2+ 1.22 Mn 2+ 0.82 Zn 0.07 (Sb 0.50 Mn 3+ 0.23 Si 0.06 Ti 0.04 Al 0.03 Fe 0.02 )O 4 . Combined X-ray and electron diffraction studies show that tegengrenite is rhombohedral, space group R3̄ or R3, and pseudocubic. The individual tegengrenite octahedra consist of eight twin domains that give rise to complex diffraction patterns and make them unsuitable for a conventional structure determination. Tegengrenite, like the chemically related mineral filipstadite, bears a close structural relationship to spinel. The eight strongest lines in the X-ray diffraction pattern are [d (in angstroms) (I/I 0 )(hkl)]: 4.98(20)(211, 003), 4.32(19)(122), 4.24(18)(113), 3.052(33)(140, 214), 2.608(100)(241, 143, 125), 2.162(28)(244), 1.665(30)(363, 075), 1.531(26)(820) and 1.527(29)(428). The refined unit-cell parameters (hexagonal setting) are a = 16.196(1) and c = 14.948(2) Å with Z = 42; for the cubic spinel-type subcell A = c/√3 - = 8.63 Å. The new species is named for Felix Tegengren (1884-1980).