Band 83, Heft 7-8

The partitioning behavior of Cr into olivine in basaltic systems has been parameterized and can now be modeled over a wide range of redox conditions and liquid compositions. The Cr 2+ /Cr 3+ in spinel-saturated experimental systems can be estimated based on a simple model of Cr solubility in basalt. Fe 3+ appears to suppress the presence of Cr 2+ in basaltic systems. We predict that, in Fe-free systems, all Cr is trivalent at log f O₂ = -3 (i.e.. QFM+3 to QFM+4). whereas all Cr is trivalent at approximately Ni-NiO(QFM+1) in Fe- bearing systems. Cr 2+ predominates under redox conditions <IW-1 in both Fe-bearing and Fe-free systems. D Cr²+ , and D Cr³+ (olivine/liquid) have been determined in various liquid compositions and temperatures. D Cr³+ (i.e., f O₂ ≥ QFM, appropriate for most terrestrial or martian basalts) strongly covaries with the ratio of non-bridging oxygens to tetrahedrally coordinated cat­ions (NBO/T) (Mysen 1983) and can be estimated using the equation This relationship appears to be valid over the entire pressure range of olivine stability, from I atm to 15 GPa. D Cr²+ (i.e., ≤ IW-1, appropriate for lunar and some asteroidal basalts) is sensitive to liquid composition and temperature and can be estimated using either or The 1/T equation is probably only valid at I atm pressure, but the D Mg equation may be useful at higher pressures as well. The Cr content of spinel-saturated liquids is a function of temperature, composition, and f O₂ . However, the Cr content of spinel-saturated liquids is buffered by spinel and is insensitive to the bulk Cr content of the system (e.g., Roeder and Reynolds 1991). Therefore, the Cr content of a crystallizing, spinel-saturated basalt cannot be modeled using Raleigh fractionation models.

We have measured multicomponent chemical diffusion coefficients in a melt near to the low pressure water-saturated eutectic granite composition in the system K 2 O-Na 2 O-Al 2 O 3 -SiO 2 -H 2 O at 1.0 GPa and temperatures of 1300 and 1600 ℃. The measured diffusion profiles can be accounted for within the analytical error by diffusion coefficients, which are not dependent on composition within the range of compositions accessed by our experiments. The multicomponent diffusion coefficient matrix [D] has a highly degenerate set of real, positive eigenvalues that show a regular relation to melt viscosity on an Ar­rhenius diagram. The smallest eigenvalue is that associated predominantly with Si-Al ex­change. The larger two eigenvalues are those associated with Si-Na and Si-K exchange and are effectively degenerate, with the result that exchanges of alkalis for silica or for each other can proceed in pseudo-binary fashion without inducing fluxes of other components. The eigenvalue associated with H-Si exchange is smaller than the alkali-silica eigenvalues, but analytical uncertainties make it also effectively degenerate with the alkalis. Uphill diffusion, notably of) water and alkalis, was observed in several experiments, and this would lead to transient partitioning of water and alkalis across diffusion interfaces showing large Al 2 O 3 concentration gradients. Such partitioning in natural systems would persist until Al concentration gradients were erased by continued, much slower Al-Si interdiffusion.

Oxygen self-diffusion in rutile was studied in synthetic and natural samples over the temperature range 750 to 1000 ℃and the pressure range 0.1 to 1000 MPa using an 18 O- enriched source. Most experiments investigated the dependence of D on temperature, water pressure, crystallographic direction, and experiment duration. A few experiments investi­gated the dependence of D on f O₂ , and confining pressure. The uptake profiles of 18 O in experimental products were measured by nuclear reaction analysis using the reaction 18 O(ρ, α) 15 N. Two mechanisms are responsible for O diffusion in rutile, and one is faster than the other by about an order of magnitude. O that diffuses by the faster mechanism is described by the diffusion law: D(∥c) = 4.7 × 10 -7 exp(-258 ± 22 × 10 3 /RT) D 0 in m 2 /s; E A in J/mol; T in K. Diffusion by the slower mechanism is described by this law: D(∥c) = 5.9 × 10 -5 exp(-330 ± 15 x 10 3 /RT) D 0 in m 2 /s; E a in J/mol:T in K. Oxygen fugacity in itself does not affect D at fugacities between I atm and Ni-NiO. However, the presence or absence of water during reduction does affect the diffusion behavior. When water is absent during rutile growth and/or subsequent reduction, only the faster mechanism operates, and when water is present during growth or reduction, both mechanisms operate simultaneously, though the contribution from the slow mechanism dominates that of the fast mechanism. Because few geologic environments are truly dry. the slower law should generally be used for modeling O diffusion for rutile in nature. Comparison with other studies of rutile suggests that migration of O vacancies is the mechanism responsible for the faster diffusion law whereas migration of Ti interstitials is responsible for the slower diffusion law. Oxygen diffusion in rutile is slower perpendicular to the c axis than parallel to that axis by about half an order of magnitude. There is no perceptible effect of confining pressure on D below 100 MPa, or between 600 and 1000 MPa. However, between 100 and 600 MPa. D decreases by nearly an order of magnitude. Closure temperatures for O diffusion in rutile are high-650 ℃ for a crystal with a 100 μm radius and a 10 ℃/Ma cooling rate. Rutile is retentive of its O isotopic composition. A crystal with a 100 μm radius will retain its initial core composition for just over 10 million years at 600 ℃.

For iron-rich compositions, reactions among titanite, clinopyroxene, olivine, quartz, magnetite, and ilmenite can constrain the conditions under which common rocks form. Phase equilibrium experiments in the system CaO-FeO-Fe 2 O 3 -TiO 2 -SiO 2 constrain two of these reactions: clinopyroxene + ilmenite = titanite + olivine (1) clinopyroxene + ilmenite = titanite + magnetite + quartz (2) in the range 600-1100 ℃ and <1 bar -3.8 kbar, and under controlled f O₂ conditions. The experimental results indicate that the assemblage clinopyroxene-ilmenite represents reduc­ing conditions relative to titanite-magnetite-quartz. However, the latter is not necessarily indicative of oxygen fugacity conditions higher than that of the FMQ buffer. The data also strongly suggest that the equilibrium positions of reaction I and 2 reflect changes in pres­sure. temperature. f O₂ ,. a SiO₂ ,. μFe 2+ Ti 4+ Fe 3+ -2 and μCaFe -1 in the coexisting Fe-Ti oxides and silicates. Standard state thermodynamic data and solution models for Ca-Fe olivine and clinopyroxene solid solutions. Fe-Ti ilmenite and magnetite solid solutions, and quartz (QUILF database) allow refinement of standard state thennochemical data for end-member titanite (A2/a space group): ΔH 0 = -2607.41 kJ/mol S 0 = 106.00 J/(mol · K) ΔG 0 = -2458.95 kJ/mol and require modification of published heat capacity equations to: The derived data for titanite are internally consistent with an enthalpy of formation of -2610.13 ± 2.90 (2 std. error) kJ/mol, established by drop solution calorimetry, and are specific to the QUILF database. However, analysis of the experimental results using dif­ferent internally consistent databases suggest that the revision proposed in this study of the currently accepted entropy value of 129.20 J/(mol · K) is largely independent of the databases used. The new entropy is required by the enthalpy of formation for titanite adopted here, which may also require revision of the heat capacity.

The equilibria: tremolite + forsterite = 2diopside + 5enstatite + H 2 O (1) tremolite + 3calcite + 2quartz = 5diopside + 3CO 2 + H 2 O (4) have been reversed experimentally at P fluid = P H₂O = 0.5 kbar, 1.0 kbar, and 5.0 kbar and at P fluid = P H₂O + P CO₂ = 5 kbar, respectively. Starting materials consisted of natural trem­olite (St. Gotthard, Switzerland) and quartz (Brazil), and synthetic calcite, forsterite, diopside, and enstatite mixed in stoichiometric proportions. Reaction direction was deter­mined by comparing XRD patterns of reactant and product assemblages and by examining surface features of experimental products with an SEM. Our new experimental data for Equilibrium 1 are consistent with the natural-tremolite results of Skippen and McKinstry (1985), who used St. Gotthard tremolite, whereas the new bracket for Equilibrium 4 is ≈25 ℃ lower than that of Slaughter et al. (1975) who also used St. Gotthard tremolite. Comparison of our results with other studies indicates that use of the St. Gotthard tremolite in place of synthetic tremolite in the starting material displaces these equilibria toward higher temperatures by about 25 and 5 ℃. respectively. Tremolite stability differences reflected in these data, as well as m phase equilibrium data for nine additional equilibria involving synthetic and natural tremolite can be accounted for with a simple ideal on-site mixing model to describe tremolite compositional differ­ences. Our analysis leads us to conclude, however, that tremolite growth in some experi­ments near the equilibrium boundary occurs with respect to metastable end-member py­roxenes used in starting materials, whereas pyroxene-stable half-brackets involve growth of stable pyroxene compositions. Thennodynamic properties for end-member tremolite, retrieved by mathematical programming analysis of the experimental phase equilibrium data with these assumptions, provide the most sound basis for prediction of calcic aniphi- bole stability relationships in natural assemblages, as well as improved calibration of quan­titative amphibole geothermobarometers. Our success hi extracting consistent thermody­namic properties for end-member tremolite from experimental data obtained with both synthetic and natural tremolite. assuming the former to contain 10 mol% magnesiocum- mingtonite component (Jenkins 1987), can be taken either as support for the validity of this assumption or as an indication that chain multiplicity faults (Maresch et al. 1994) produce a similar degree of stabilization as this solid solution.

New values are reported for the Gibbs energies of formation from the elements for hydrocerussite Pb(OH) 2 · (PbCO 3 ) 2 and hydrozincite [Zn(OH) 2 ] 3 · (ZnCO 3 ) 2 . These ΔG 0 f values were obtained from electrochemical cells without liquid junction. We determined ΔG 0 f [Pb(OH) 2 · (PbCO 3 ) 2 (S)] = -1699.8 ± 1.6 kJ/mol for hydrocerussite and ΔG 0 f ([Zn(OH) 2 ] 3 · (ZnCO 3 ) 2 ) = -3163.3 ±4 kJ/mol for hydrozincite. These results allow future electrochemical cell experiments to be performed to determine the ΔG 0 f values of other hydroxycarbonate minerals using either the Pb amalgam-hydrocerussite or the Zn amalgam-hydrozincite as reference electrodes. These reference electrodes provide a strategy for establishing Gibbs energies for phases with two different anions, which are geochemically interesting but difficult to study experimentally.

The solubilities of analcime and clinoptilolite were determined in dilute, weakly alkaline, aqueous solutions below 300 ℃ and at vapor-saturated pressures. Analcimes used in this study were from Mont St. Hilaire. Quebec (an 1 , Si/Al = 2.02) and Wikieup, Arizona (an 2 , Si/Al = 2.55): clinoptilolite samples were from Castle Creek. Idaho (Si/Al = 4.50). The effects of alkali content (Na,K) on clinoptilolite solubility were determined by using cation-exchanged varieties of the Castle Creek material (cp 1 , cp 2 ). In neutral to weakly alkaline solutions, the dominant solubility-controlling reactions of these zeolites are NaAlSi 2 O 6 · H 2 O(an 1 ) + 5H 2 O(1) ↔ Na + + Al(OH) - 4 + 2Si(OH) 4 (aq), Na 0.85 Al 0.85 Si 2.15 O 6 · H 2 O(an 2 ) + 5H 2 O(1) ↔ 0.85Na + + 0.85Al(OH) - 4 + 2.15Si(OH) 4 (aq), Na 1.1 Al 1.1 Si 4.9 O 12 · 3.5H 2 O(cp 1 ) + 8.5H 2 O(1) ↔ 1.1Na + + 1.1Al(OH) - 4 + 4.9Si(OH) 4 (aq), and K 1.1 Al 1.1 Si 49 O 12 · 2.7H 2 O(cp 2 ) + 9.3H 2 O(1) ↔ 1.1K + + 1.1Al(OH) - 4 + 4.9Si(OH) 4 (aq). The logarithm of the equilibrium constants of these reactions were fit to the function: log K = A + BT + C/T + D log T. At 25 ℃, log K 25 values for the Mont St. Hilaire analcime, Wikieup analcime, Na-clinoptilolite, and K-chnoptilolite are -16.1, -15.0, -26.5, and -28.1, respectively. These data were combined with the thermodynamic properties of the aqueous (aq) species Si(OH) 4 , Al(OH) - 4 , Na + , K + , and liquid water (1) to determine standard Gibbs free energies of formation as a function of temperature. Values of ΔG 0 f at 25 ℃ and 1 bar for the Mont St. Hilaire analcime and Wikieup analcime are -3089.2 and -3044.4 kJ/mol. respectively. The ΔG 0 f values for hydrous Na-clinoptilolite and K-clinoptilolite, respectively, are -6267.9 and -6107.4 kJ/mol at 25 ℃ and 1 bar. The solubility data reported here, and results obtained from previous calorimetric studies, indicate that the aluminosilicate frameworks of analcime and clinoptilolite are stabilized by an increase in Al content.

We studied mixed-layer illite-smectite (I/S) (<1 or 0.5 μm size fractions) in 28 diagenetic bentonite samples, Pliocene to Ordovician in age, from different locations in North America and Europe to investigate the smectite-to-illite transformation mechanism. XRD- measured illite contents ranged from 3 to 100%, and layer ordering varied continuously from R = 0 to R ≥ 3. XRD polytype analysis showed that with progressive illitization, ES changes from a predominantly cis-vacant to a trans-vacant structure, although considerable scatter is present in the data. There is no evolution of illite in the I/S from IM d to 1M polytypes with progressive illitization. Microprobe chemical analysis and combustion analysis for N showed that NH + 4 comprised 9-22% of the total fixed cation content. We calculated an interlayer charge for the illite end-member of 1.5 per O 20 (OH 4 ). Sample morphology was studied by atomic force microscopy, which showed a constant aspect ratio and particle diameter over the whole I/S series. Particle thickness increased with percent illite. Oxygen isotope compositions of the I/S clays did not display any systematic trends. Our results suggest that illitization of bentonite I/S proceeds via a solid-state mech­anism, with local lattice rearrangements that permit structural (octahedral cis/trans occu­pancy) and chemical changes in 2:1 layers and cause interlayer coalescence.

Phengitic muscovite-2M 1 crystals [ [12] (K 0 . 88-0 . 99 Na 0 . 01-0.09 Ca 0 . 00-0.06 Ba 0 . 00-0.01 ) [6] (Al 1 . 64-1.88 Fe 2+ 0.06-0 . 29 Fe 3+ 0.01-0.16 Mg 0.00-0.16 Mn 0.00-0.07 Ti 0.00-0.06 ) [4] (Si 2.87-3.30 Al 0.70-1.13 )(OH) 1.56-2.07 F 0.00-0.41 O 9.91-10.25 ] from pegmatites and peraluminous granites were refined to investigate the influence of phengitic substitution on the mica structure. Single-crystal X-ray diffraction data were collected for eleven crystals in the C2/c space-group (agreement factor 2.1% ≤ R obs ≤ 3.9%). Tetrahedral Si and Al cation disorder was found for each sample, with the mean tetrahedral cation-oxygen distances ranging from 1.639 Å ≤ <T1-O> ≤ 1.647 Å and 1.640 Å ≤ <T2-O> ≤ 1.646 Å. As phengitic substitution increases, the octahedral sheet expands and requires a less distorted (more hexagonal) tetrahedral ring (7.70° ≤ α ≤ 11.38°) and low corrugation of the basal O plane (0.1796 Å ≤ Δz ≤ 0.2296 Å). The electron density at the M2 site is greater than that required for the ideal muscovite-2M 1 structure, and a small excess of electron density is found in the M1 site. The inner sixfold coordination of the interlayer (A) cation is elongated along c*, which is consistent with the high α values and the long A-O11 bond length.

The occurrence of hausmannite with an apparent Curie temperature close to 750 K, instead of 41.8 K was recently described from hydrothermally altered manganese ore from the Kalahari manganese field, South Africa. The unusual magnetic properties were related to the substitution of Fe 3+ for Mn in the hausmannite structure. Because of the large differences in the scattering lengths of Fe and Mn, b Fe = 9.94 and b Mn = -3.73 fm, respectively, we performed neutron powder diffraction experiments at 295 and 10 K on natural mineral separates and synthetic compounds to determine the influence of the Fe substitution on the crystal structure and the magnetic properties of the hausmannite. Riet- veld refinements of synthetic Fe-rich hausmannite neutron powder diffraction patterns at 295, 60, and 10 K indicate some significant and interesting changes of magnetic properties and crystal structure of hausmannite, which are directly linked to an increasing amount of iron substituting for manganese. The unit-cell parameters of Mn 3-x Fe x O 4 , in particular, il­lustrate decreasing Jahn-Teller distortion with increasing Fe content, whereas the Curie temperature was found to increase significantly with increasing Fe content. Nevertheless, this study indicates that the presence of Fe-rich hausmannite causes the unusual high-temperature ferrimagnetic behavior in the Kalahari manganese field.

We present a new oxybarometer for magnesiowüstite-bearing systems, which is easily applied using widely available techniques. A scale relating the proportion of Fe 3+ [α = Fe 3+ /(Fe 3+ + Fe 2+ )] to the position of the (220) reflection and total Fe (y) in magnesiowüstites (Mg 1 -y · Fe y ) 1-x O has been derived from measured values in samples equilibrated at various oxygen fugacities: α = 13.0047 - 39.4829f + 40.0540f 2 - 13.5701f 3 (±0.007 + 0.09α), f = (d 220 - 1.4890)/(0.0510y + 0.0206y 2 )(±0.0001). Previously established partition coefficients can then be used to relate estimated Fe 3+ content to equilibrium oxygen fugacity (f O₂ ) in the composition range 0 < y < 0.2. Equi­librium oxygen fugacities of log f O₂ ≤ -1.7 can be estimated to ±0.5 log units using just X-ray powder diffraction and electron microprobe analytical techniques.

The Fe, Mg, and vacancy distributions in oxidized olivine of intermediate composition and the crystal structures of associated planar defects were examined by a combination of atomic-resolution transmission electron microscopy, image analysis, and image simulation. Simulated images using various structures and site occupancies show that the defects are not laihunite-like layers within an essentially normal olivine, but instead are integral parts of a highly oxidized crystal with a high degree of cation and vacancy ordering and a distorted olivine structure. A possible new structure was derived as a variant of previously published refinements for oxidized fayalite (laihunite-3M), with one defect row per unit cell instead of two. Details in high-resolution images are extremely sensitive to variations in atomic positions and site occupancies. Semiquantitative constraints on relative magni­tudes of electron densities provided by image simulation were combined with crystal chem­ical constraints based on local charge balance and compositional constraints from analytical electron microscopy in a series of linear equations, which were solved to provide several candidate models for the distribution of Fe, Mg, and vacancies. Simulations using the new structure and an occupancy model in which essentially all of the Fe is Fe 3+ in M2 sites and all of the vacancies are in Ml sites match significant characteristics of HRTEM images obtained at Scherzer defocus and at overfocused conditions. This ordering scheme is sim­ilar to that of true laihunites and may result from differences in site and cation sizes.

Transformation twinning in minerals forms isolated twin walls, intersecting twin walls with comer junctions, and wedge-shaped twins as elements of hierarchical patterns. When cut perpendicular to the twin walls, the twins have characteristic shapes, right-angled and needle-shaped wall traces, which can be observed by transmission electron microscopy or by optical microscopy. Theoretical geometries of wall shapes recently derived for strain- related systems should hold for most displacive and order-disorder type phase transitions: (I) right-angled twins show curved junctions; (2) needle-shaped twins contain flat wall segments near the needle tip if the elastic behavior of the mineral is dominated by its anisotropy; (3) additional bending forces and pinning effects lead to curved walls near the junction that make the needle tip appear more blunt. Experimental studies confirmed that these features occur in a wide range of materials. Bent right-angled twins were analyzed in Gd 2 (MoO 4 ) 3 . Linear needle tips were found in WO 3 , [N(CH 3 ) 4 ] 2 · ZnBr 4 CrAl, BiVO 4 , GdBa 2 Cu 3 O 7 , and PbZrO 3 . Parabolic tips occur in K 2 Ba(NO 2 ) 4 , and GeTe whereas exponential curvatures appear in BaTiO 3 , KSCN, Pb 3 (PO 4 ) 2 , CaTiO 3 , alkali feldspars, YBa 2 Cu 3 O 7 , and MnAl. The size and shape of the twin microstructure relates to its formation during the phase transition and the subsequent annealing history. The mobility of the twin walls after formation depends not only on the thermal activation but also on the structure of the wall, which may be pinned to impurities on a favorable structural site. Depinning energies are often large compared with thermal energies for diffusion. This leads to kinetic time scales for twin coarsening that are comparable to geological time scales. Therefore, transformation twins that exhibit needle domains not only indicate that the mineral underwent a structural phase transition but also contain information about its subsequent geological history.

The structure of sodium tetrasilicate (Na 2 Si 4 O 9 ) glasses containing 0 to 10 wt% water was investigated by a combination of Raman, IR, and NMR methods. Both the 29 Si magic angle spinning NMR data and Raman spectra in the Si-O stretching region clearly show that water depolymerizes the silicate network of the glasses. Q-species distributions cal­culated from Raman spectra, assuming equal scattering cross sections of all bands in the Si-O stretching region, closely agree with results obtained from NMR data. At low total water contents, the silicate network is depolymerized mainly by breaking of Q 4 -Q 4 bonds, whereas breaking of Q 3 -Q 3 bonds dominates at high water contents. Near IR spectra show the presence of both OH groups and molecular H 2 O in the glasses. The number of non­bridging O atoms per silicon atom, calculated from the near IR data, closely agrees with the results obtained from Raman and NMR, and confirms the assignment of the 4500 cm -1 band in the near IR to a combination mode of Si-OH groups. Moreover, the intensity of the fundamental Si-OH stretching band at 910 cm -1 in the Raman spectra varies propor­tionally to the intensity of the 4500 cm -1 near IR band. Both IR and Raman spectra show three main bands in the OH-stretching region, centered at 3580, 3000, and 2350 cm -1 , due to hydrous species with different hydrogen bond strengths. The relative intensities of these three bands are insensitive to total water content and 0H/H 2 0 ratio, suggesting that both OH and H 2 O contribute to each of these bands. This is consistent with the fine structure of the H 2 O bending vibration in the IR spectra around 1640 cm -1 and with the polarization dependence of the OH-stretching bands in the Raman spectra. Near IR spectra of hydrous sodium tetrasilicate glasses and hydrous aluminosilicate glasses are very similar and show a similar dependence of band intensity on total water content, suggesting that there is no fundamental difference in the dissolution mechanism of water in these systems.

Twenty-two andradite samples from a variety of geological environments and two synthetic hydroandradite samples were studied by Fourier transform IR spectroscopy. Their spectra show that H enters andradite in the form of OH - . Amounts up to 6 wt% H 2 O occur in these samples; those from low-temperature formations contain the most OH - . Some features in the absorption spectra indicate the hydrogamet substitution (SiO 4 ) 4- ↔(O 4 H 4 ) 4- whereas others indicate additional types of OH - incorporation. The complexity of the spectra due to multi-site distribution of OH - increases with increasing complexity of the garnet composition.

The lattice dynamics computer code PARAPOCS was successfully used to calculate the 240 vibrational frequencies of pyrope garnet, Mg 3 Al 2 Si 3 O 12 , at ambient conditions. The atomic displacement vectors (eigenvectors) for each frequency were also calculated and their symmetry relations analyzed with the aid of factor group analysis (FGA), to determine the symmetry species of each vibrational mode. Comparison with the experimental IR and Raman data shows excellent agreement, but no LO-TO reversals were identified. Calcu­lation of the frequency shifts due to the isotopic substitution of 26 Mg and 30 Si, together with a more detailed analysis of the calculated eigenvectors, enabled identification of the dominant site or cation motion contributing to each vibrational mode. Previous assignments of the high-frequency vibrations to pure SiO 4 internal modes and the lower-frequency vibrations to mixed cation modes are supported. We conclude that the specific number of site/atom motions predicted by site group analysis (SGA) is not adhered to due to substantial mode mixing, and that FGA and SGA, in which the SiO 4 tetrahedra are treated as isolated units, are only applicable at high frequencies. The agreement observed between the calculated and experimental data leads us to conclude that the method of computer modeling used and the interatomic potentials employed in the simulations provide a good description of the lattice dynamical behavior of pyrope garnet.

We report here the results of a series of inclusive chemical characterizations, including all elements except oxygen, for a suite of 54 tourmaline samples. A combination of analytical techniques was used to analyze for major and light elements (electron microprobe), Fe 3 + and Fe 2 + (Mossbauer spectroscopy), H (U extraction), and B, Li, and F (ion microprobe, or SIMS). The B content of the tourmalines studied ranges from 2.86 to 3.26 B per formula unit (pfu) with 31 anions; excess boron is believed to reside in the Si site. Li ranges from 0.0 to 1.44 Li pfu and F contents are 0.0-0.91 pfu. H contents range from nearly anhydrous up to 3.76 H pfu and do not correlate simply with Fe 3 + content. Mössbauer results show that tourmaline exhibits the entire range of Fe 3+ ∑Fe from 0.0-1.0. Fe 2+ is represented in the spectra by three doublets, with occupancy in at least three distinct types of Y sites (with different types of nearest and next nearest neighbors). Fe 3 + was found in 26 of the 54 samples studied. Although Mössbauer data do not allow the distinction between [Y] Fe 3+ and [Z] Fe 3+ site occupancies to be made, XRD data on these samples suggest that the majority of Fe 3 + is also in Y. Of the samples studied, [4] Fe 3+ occurs in nine; five of those were either olenite or uvite with extensive Na substitution. A mixed valence doublet corresponding to delocalized electrons shared between adjacent octahedra was observed in 14 of the samples studied. Projection pursuit regression analysis shows that distribution of Fe among doublets is a function (albeit a complex one) of bulk composition of the tourmaline and supports the interpretation of doublets representing different populations of neighbors. Variations in Fe 3+ / Fe 2+ ratio cannot be directly related to variations in charge in any single site of the structure. Fe 3+ /Fe 2+ ratio is probably controlled by the prevailing oxidation state in the bulk rock assem­blage, rather than by any particular crystal chemical substitution.

Optical band gaps for a set of 23 ternary and quaternary sulfide minerals have been measured by diffuse reflectance spectroscopy. Comparison of band gaps measured by diffuse reflectance with band gaps determined by single-crystal methods for 12 binary sulfides demonstrates that the diffuse reflectance measurement produces results accurate to within 0.1 eV. Unlike the band gaps of binary sulfides that plot linearly with various measures of bond energy, the ternary and quaternary sulfides cluster within ranges determined by the chemical composition. The band gaps for the ternary and quaternary sulfides tend to lie between the band gaps of the component binary sulfides.

Incommensurate tridymite L3-T O (MX-1) shows a cascade of five phase transitions at 65, HO, 150, 200, and 380 ℃ upon heating. The X-ray diffraction patterns were investigated with a Buerger precession camera revealing a sequence of four incommensurate phases in the range from room temperature to 200 ℃. The phases formed between 65 and HO ℃ and between HO and 150 ℃ are new modifications of tridymite. At 65 ℃ the monoclinic tridymite L3-T 0 (MX-1) phase undergoes a first-order transformation to an orthorhombic phase. The incommensurate structural modulation of the room-temperature phase with the wavevector q x = 0.663 a* -0.498 c* flips to q 2 = 0.042 a* -0.388 c*. Simultaneously, a commensurate modulation with tripled b lattice parameter is formed. The wavelengths of both modulations do not depend appreciably on the temperature. The incommensurate modulation discontinuously disappears near 110 ℃ whereas the commensurate modulation along the b axis becomes non-integral with a temperature-dependent wavelength varying between 115 and 100 A. At 150 Å the symmetry is reduced to monoclinic again with 7 = 90.4°. Between 150 and 200 ℃ the monoclinic angle gradually decreases to 90° and the wavelength of the modulation from about 90 to 65 Å. At higher temperatures, the satellite reflections fade into weak streaks and the normal orthorhombic high-temperature modification of tridymite is formed. The phase transitions are reversible upon cooling except for the first transformation that is partly irreversible for single crystals and reversible but incomplete for pulverized material at room temperature.

Phase egg, first described by Eggleton et al. (1978), was synthesized and its composition determined to be AlSiO 3 OH. The crystal structure of AlSiO 3 OH, including the position of the hydrogen, has been solved and refined from high-resolution X-ray powder diffraction. The resulting lattice constants are a = 7.14409(2) Å, b = 4.33462(1) Å, c = 6.95253(2) Å, and β = 98.396(1)°. The space group is P2 1 \n; Z = 4, V 0 = 212.99(1) Å 3 , and the zero pressure density is 3.74 g/cm 3 . This phase, which has features in common with the sti-shovite structure, occurs above 11 GPa and 700 ℃. AlSiO 3 OH forms from topaz-OH with increasing pressure and persists to more than 17.7 GPa and 1300 ℃. From a Schreinemaker analysis, we predicted that phase egg decomposes with pressure to an unknown, possibly hydrous aluminosilicate. Potentially, phase egg could replace topaz-OH or kyanite in sub­ducted crustal bulk compositions and may transport some water into the deep Earth.

Phosphovanadylite, whose simplified formula is (Ba,Ca,K,Na) x [(V,Al) 4 P 2 (O,OH) 16 ] · 12H 2 O, is a new vanadium phosphate zeolite mineral found in the Phosphoria Formation at Monsanto’s Enoch Valley Mine, Soda Springs, Idaho. Its formula in more detail is (Ba 0.38 Ca 0.20 K 0.06 Na 0.02 ) ∑0.66 [P 2 (V 3.44 Al 0.46 ) ∑0.39 O 10.34 (OH) 5.66 ]·12H 2 O. The drusy mineral occurs as pale greenish-blue euhedral cubes (20-50 μm edge) coating phosphatic, organic-rich mudstone. The chemical composition determined by electron microprobe is (in weight percent) V-28.02, P-9.91, Al-1.97, Ca-1.31, Ba-8.28, Cd-0.09, Zn-0.34, Na-0.15, K-0.73, O-46.57, and F-0.03. The index of refraction is n D = 1.566 (4) and specific gravity is 2.16 (3). The X-ray powder pattern shows strong reflections at 3.16 Å (422), 2.58 (600), 2.44 (620), and 7.73 (200), which are indexed on the basis of a cubic body-centered unit cell with a = 15.470 (4) Å. From the single-crystal structure analysis, its space group was determined to be I4̅3m, Z = 6, and its structure consists of V 4 O 16 octahedral clusters linked to each other by P atoms to form a cubic lattice, creating cavities 7.0 and 5.5 Å in diameter where mainly H 2 O resides. Final residual indexes are R = 0.066, R w = 0.061, goodness- of-fit = 0.75, and 93 observations and 24 parameters.

Rossmanite is a new tourmaline species from near Rožná, western Moravia, Czech Republic. It forms pale pink columnar crystals about 25 mm long and 5 mm thick, elongaten along c with striations parallel to c on the prism faces. It is brittle, H= 7, D meas = 3.00 g/cm 3 , D calc = 3.06 g/cm 3 . In plane-polarized light, it is colorless. Rossmanite is uniaxial negative, to = 1.645(1), e = 1.624(1), trigonal, space group R3m, in the hexagonal setting a = 15.770(2), c = 7.085(1) Å, V = 1525.8(4) Å 3 , Z= 3. The strongest six X-ray diffraction lines in the powder pattern are at d = 3.950 Å with I = 100% for (hkl) = (220); 2.552 Å, 93%, (051); 1.898 Å, 72%, (342); 4.181 Å, 58%, (211); 2.924 Å, 56%, (122); and 3.434 Å, 53%, (012). Analysis by a combination of electron microprobe, SIMS, H-line extraction, and crystal-structure refinement gave SiO 2 38.10 wt%, Al 2 O 3 44.60, Na 2 O 1.43, Li 2 O = 1.13, B 2 O 3 = 10.88, H 2 O = 3.70, F = 0.20, O ≡ F 0.08, sum = 99.96 wt%, Fe, Mg, Ca, Mn, Ti, F, K not detected. The formula unit (31 anions) is x (⃞ 57 Na 0.43 ) Y (Li 0.71 Al 2.17 ) z Al 6 (Si 5.92 O 18 ) (B 2.92 O 9 )(OH) 3.83 F 0 . 10 O 0.07 , with the ideal end-member formula ⃞(LiAl 2 )Al 6 (Si 6 O 18 )(BO 3 ) 3 (OH) 4 ; thus rossmanite can be derived from elbaite [Na(Al 1.5 Li 1.5 )(Si 6 O 18 )(BO 3 ) 3 (OH) 4 ] by the substitution x ⃞ 2 + Y Al → x Na 2 + Y Li, where ⃞ = vacancy. The crystal structure of rossmanite was refined to an R index of 1.7% using 1094 observed (5σ) reflections collected with MoKα X-radiation from a single crystal. The structure refinement confirmed the low occupancy of the X site and the presence of Li at the Y site. There is considerable positional disorder at the O1 and O2 sites induced by the local variations in bond-valence distribution associated with ⃞-Na disorder at X and Li-Al disorder at Y.

Sorosite, ideally Cu(Sn 5 Sb), is a new mineral species from the Baimka gold-platinum- group mineral placer deposit, Chukotka, Russian Far East. It occurs as large subhedral to euhedral crystals (0.1-0.4 mm in length), hexagonal in cross section, minute crystals (≤ 15 μm, also hexagonal), and anhedral grains. Sorosite forms inclusions in Sb-bearing native tin, is often intergrown with stistaite (Sn 1.12 - 1.13 Sb 0.87 - 0.88 ) and occurs with herzenbergite (SnS), native lead, and trace cassiterite. Sorosite is brittle with a microhardness VHN 40;50 = 443.7 kg/mm 2 (n = 3). No cleavage is observed. In reflected light, the large crystals are nearly white with a pinkish tint, whereas the microcrystals show a pronounced pinkish tint. Bireflectance is variable. The average of nine electron microprobe analyses gave Cu 35.33, Fe 1.18, Sn 58.18, and Sb 4.77, sum 99.46 wt%, corresponding to (Cu 1.00 Fe 0.04 ) ∑1.04 (Sn 0.89 Sb 0.07 ) ∑0.96 . The powder pattern is close to those of natural Cu(Sn,Sb) and synthetic η-Cu 6 Sn 5 ; it was indexed for a hexagonal cell, with a = 4.217(4) Å, c = 5.120(6) Å, and V = 78.85 Å 3 . For Z = 2, the calculated density is 7.6 g/cm 3 . The strongest lines in the pattern are at 2.970 (011), 2.112 (110), and 2.094 Å (012). The sorosite-bearing mineral assemblage apparently formed under low f O₂ and f S₂ conditions.

Time-of-flight secondary ion mass spectrometry (TOF-SIMS) yields mass spectra of the upper several monolayers of an analytical surface. The applicability of TOF-SIMS to the characterization of C compounds on crack surfaces in rocks is demonstrated. A pyroxene grain recovered from the interior of a gabbro xenolith from the 1801 flow of Haulalai volcano, Hawaii, was chosen for this initial study because well-developed carbonaceous films are known to exist on many of the crack surfaces in these rocks. In addition to the anticipated major elements (Si, Al, Fe, Mg, Ca), several minor and trace elements (B, Li, Na, Ti, Mn, Co, Ni, Cu) were identified in the positive ion mass spectra. The unsputtered surface is covered with a hydrocarbon-rich layer, as indicated by the presence of numerous light C x H y fragments as well as several masses of several hundred atomic mass units (amu). This layer is much better developed than the ubiquitous atmosphere-derived hydrocarbon layer observed on nominally clean, unsputtered surfaces and therefore must be indigenous. High concentrations of Ni and Cu are associated with the C-rich layer and may exist as organo-metallic compounds. Several C-N fragments, possibly indicative of biogenic com­pounds, were also identified in negative ion spectra. Imaging reveals the presence of localized islands enriched in oxides of Si, Al, Na, and Ca beneath the carbonaceous layer. This study demonstrates that TOF-SIMS can be used to provide unique and geochemically useful information on crack surfaces in rocks.

This paper demonstrates that the six standard polytypes of chlorite, whose definitions are based on the orientation of the interlayer sheet and the position of the interlayer sheet on the 2:1 layer, can be discriminated by atomic-resolution images recorded down [010], using a transmission electron microscope with a ~2 Å point resolution and digital image processing. Several specimens were investigated to reveal their local stacking structures. A Ibb chlorite is highly twinned and twin boundaries consist of a IIb+Ia stacking sequence at the interlayer sheet. An interstratified chlorite/biotite formed by hydrothermal alteration from biotite in granite consists of a mixture of several chlorite polytypic sequences, including Iab, Ibb, IIab, and IIbb. These polytypic details of chlorite and other sheet silicates provide important insights into mineral stability, origin, and reaction mechanisms.

Significant advances have been made in hydrofluoric acid solution calorimetry at Lafayette College in the past 15 years. To determine the degree to which these developments enable the reduction of sample size, calorimetric experiments were performed on hexagonal germanium oxide as a function of sample weight. The resulting calorimetric data indicate that the highest degrees of reproducibility (±0.1%) are maintained down to sample sizes of 50 mg, and that precisions of ±1%, acceptable for many applications, are observed to sample sizes of 10 mg. Because silicate systems produce weight-based heats of solution that are about twice that of germanium oxide, the required sample size for these will be even less. The new minimum required sample size of 5 to 25 mg (depending on applica­tion) is about two orders of magnitude less than that used 20 or 30 years ago. This makes possible many new kinds of projects for HF solution calorimetric investigation, including those on high-pressure materials.