Volume 81, Issue 5-6

Igneous cumulate rocks provide an important record of planetary magmatism, but there are pitfalls in their interpretation. Cumulus minerals may react with the trapped melt and other cumulus phases during sub solidus reactions, thus losing a direct record of their igneous history. One of the best approaches for estimating the melt compositions parental to the cumulates is to analyze the cores of cumulus phases for elements with slow diffusion rates (e.g., REE in pyroxene) because these most reliably retain a record of the mineralmelt partitioning. We have conducted SIMS studies of pyroxenes from a variety of planetary cumulates, including lunar norites, martian orthopyroxenites, asteroidal orthopyroxenites (diogenites), asteroidal pyroxene-plagioclase cumulates (cumulate eucrites), and terrestrial orthopyroxenites and norites (Stillwater Complex, Montana). We emphasized the REE (La, Ce, Nd, Sm, Eu, Dy, Er, Yb) along with Sr, Y, and Zr in these investigations. Our studies yielded the several conclusions. (I) Lunar Mg-suite norites crystallized from highly evolved (KREEPy) melts that were emplaced into the lunar anorthositic crust. One viable model suggests remobilization of KREEP lithologies that formed late in the crystallization of the lunar magma ocean. (2) Orthopyroxenites from asteroid 4 Vesta (diogenites) likely formed as cumulates from melts derived from depleted mantle, which had previously experienced eucritic basalt removal. The calculated parental melts show a limited range of major element compositions but an exceedingly large range of trace element variation that is difficult to explain by any simple crystallization or melting models. (3) The diogenites of 4 Vesta are cumulates from melts derived from many depleted mantle reservoirs that either formed numerous orthopyroxene plutons, later mixed by impact brecciation, or commingled by magma mixing to form a limited number of differentiated plutons. We base this conclusion on a comparative investigation of a well-studied terrestrial orthopyroxenite sequence (Bronzitite zone of the Stillwater Complex). An important lesson to be learned from these studies in comparative planetology is that if simple petrogenetic models do not work for well-studied terrestrial occurrences there is little reason to believe that they will work for planetary environments where we have little geologic control on sampling.

Ab initio band-structure calculations based on the density functional theory have been performed for forsterite to obtain, with a parameter-free model, electric-field gradients for all nuclei. Calculations based on the generalized gradient approximation yield the ratio of the largest components of the calculated electric-field gradients within 1% of the experimental value for 25 Mg and within 2% of the experimental value for 17 O. The absolute values differ by about 5%, depending on which nuclear quadrupole moment is used in the conversion. The asymmetry parameters are also in good agreement with experimental data. Values obtained with a gradient-corrected exchange-correlation potential are better than those based on the standard local-density approximation. The calculated angles between the principal axes of the quadrupole coupling tensor and the crystallographic axes agree very well with the experimental data. For the Ml site the maximum deviation is 1.8°, and for the M2 site the maximum deviation is 0.6°. The current calculations allow an evaluation of three sets of experimental data for the 170 electric-field gradient. They also confirm a proposed assignment of the measured electric-field gradient tensors to specific O atoms.

The structural phase transitions of monoclinic tridymite that occur at elevated temperatures, as well as the structure of the high-temperature modifications, have been studied by 29 SiMAS NMR. The sample, originating from the surface of a refractory silica brick, displays sharp discontinuities in the first moment (M 1 ), second moment (M 2 ), or both, of the 29 Si MAS NMR spectrum at the following phase transitions: 108 ± 1 °C for the monoclinic (MC) to orthorhombic (OP) transition, 165 ± 5°C for the transition from OP to a second orthorhombic phase (OS), and 206 ± 4 °C for the transition from OS to a third orthorhombic phase (OC). M 1 for the OC phase becomes more negative with increasing temperature up to ~323°C, at which there is also a small discontinuity in M 2 , indicating a transition from OC to a hexagonal phase (LHP). The spectrum of the MC phase consists of 12 narrow lines of equal area, consistent with the known crystal structure. The spectrum of the OP phase is broad and asymmetric, indicating a distribution oflocal environments at the Si sites. It can be simulated assuming an incommensurate structural modulation with a maximum amplitude of ~5 × 10 -2 Å, but it is also consistent with a recent domain-formation model, provided that most disorder occurs at the Si6 site. The spectrum of the OS phase is asymmetric, indicating a nonlinear incommensurate modulation with a maximum amplitude of ~2 × 10 -2 Å. The spectra of higher temperature phases are consistent with the single Si site of the known crystal structures, but M 1 values imply significantly smaller mean Si-O-Si angles ‹α›, e.g., ‹α› = 152.6° at 220°C from NMR, compared with 168.0° from X-ray diffraction data. This difference is consistent with disorder of O atoms around Si-Si axes.

A diversity of twinning and exsolution textures has been observed by transmission electron microscopy in natural kalsilite from volcanic and metamorphic rocks (Mount Nyiragongo, Zaire; Alban Hills, Italy; Kerala, southern India). The phase transitions responsible for these microstructures were examined by high-temperature powder and singlecrystal X-ray diffraction. Pure kalsilite has P6 3 mc symmetry at high temperatures but transforms to its P6 3 state through a two-phase field, between ~870 and ~920°C. In this two-phase field, the low form coexists with a structure that has a sixfold repeat of its a unit-cell dimension. For ~K S₈₈ ,the high-temperature state is high tetrakalsilite with possible space group P6 3 mc; a two-phase field between ~890 and ~930°C involves the coexistence of this phase with P6 3 kalsilite. High tetrakalsilite with composition ~KS74 reverts to low tetrakalsilite on quenching to room temperature from 950°C. Latticeparameter variations, and estimates of the AI-O-Si angles that can be derived from them, imply that the stability limit for the high structural states occurs when the angles for apical and basal 0 atoms converge. At lower temperatures, pure P6 3 kalsilite appears to have transformed during a metamorphic evolution in nature to an intergrowth of the P6 3 structure and a structure for which P31c symmetry is proposed. The latter can be thought of as a polytype of kalsilite, with (001) layers stacked in an eclipsed array rather than in the staggered array of normal low kalsilite. In this regard, KAlSiO 4 seems to be remarkably similar to KLiSO 4 . If ~3-12 mol% NaAlSiO 4 is present in solid solution, a P6 3 structure with an a parameter √3 greater than normal low kalsilite develops. An anomaly in the lattice parameters of K S₈₈ suggests that the transition temperature below which this superstructure develops may be ~500°C. Natural nepheline ex solved from kalsilite displays merohedral twinning, which can be accounted for by a P6 3 mc → P6 3 transition during cooling. Transformation behavior in the Ne-Ks system may be explained, at least qualitatively, in terms of Na-K ordering between cavity sites, ordering of basal O atoms between two sites on either side of the mirror plane parallel to the c axis of P6 3 mc structures, ordering of apical O atoms between three sites, and coupling between all these processes.

We performed in-situ Raman spectroscopy and X-ray diffraction experiments at high temperature and ambient pressure to investigate the intrinsic anharmonic properties of β-Mg 2 SiO 4 and the mechanism and kinetics of its back-transformation to forsterite. Hightemperature Raman spectra of β-Mg 2 SiO 4 and its back-transformed products were recorded up to 1200 K. β-Mg 2 SiO 4 persists metastably up to 800-900 K, and the Raman frequency shifts with temperature were determined. Between 800 and 1000 K, new peaks are observed at about 670 and 1020 cm -1 . Above 1000 K, a direct transformation to forsterite occurs. The peaks that appear between 800 and 1000 K are attributed to a defective spinelloid that forms as an intermediate phase during the back-transformation of β-Mg 2 SiO 4 to forsterite. Similar features are observed in the Raman spectrum of partially transformed γ-Ni 2 SiO 4 heated at 1073 K and ambient pressure for 10 min. These results indicate that a two-step mechanism, possibly martensitic, is operative in the backtransformation of the β-and γ-phases to olivine at low to moderate temperatures and for a large overstepping of the equilibrium conditions. The kinetics of the β-to α-Mg 2 SO 0 4 back-transformation were monitored between 1023 and 1120 K at ambient pressure using X-ray powder diffraction. For the kinetic data obtained in air, two regimes are evident from an Avrami analysis. The first regime is characterized by an exponent n ≈ 2 for a low transformed fraction (X < 0.5); the second has n ≈ 1 for higher transformed fractions. For this second regime, an activation energy of 432 ± 64 kl/mol is derived for the growth process from the kinetic data. A smaller data set collected in vacuum indicates much slower transformation rates and suggests a significant effect of the O 2 or H 2 O partial pressures on the kinetics. Intrinsic mode anharmonic parameters can be calculated from the Raman frequency shifts with temperature and used to correct vibrational heat capacity models for intrinsic anharmonic effects. These corrections are slightly higher for the β-phase than for forsterite but the difference is within the experimental error. This indicates that, within the resolution of our experiments, no significant effect of intrinsic anharmonicity on the location and slope of the α-β phase transition can be predicted.

Synchrotron radiation, high-resolution transmission electron microscopy (HRTEM), X-ray powder diffraction, and Raman spectroscopy were used to study the structure and thermal behavior of malayaite, CaSnSiO 5 . No indications of deviation from A2/a symmetry and no structural transitions were observed between 100 and 870 K. HRTEM revealed that the material is free of domains and antiphase boundaries. However, the lattice constants, cell volume, and Raman-active phonons show a thermal discontinuity near 500 K, which is possibly related to variation of the coordination sphere around the highly anisotropic Ca position.

The present study attempts to clarify the structural role of Mn in two pumpellyite specimens with the chemical formulas Ca 7.80 (Al 8.91 Mn 1.85 Mg 1.16 Fe 0.37 )Si 11.84 O 56-x OH x and Ca 8.12 (Al 7.99 Mn 3.41 Mg 0.71 Fe 0.30 )Si 11.84 O 56-x OH x . The samples are from gabbroic breccias of the Bracco ophiolites (Eastern Ligurian Apennines, Italy), and their chemistry is typical of Mn-rich pumpellyite minerals associated with Mn mineralization in prehnite-pumpellyite metamorphic facies. The distribution of the Mn cations over the two crystallographically independent X and Y octahedral sites, and their prevalent oxidation states, were determined by Rietveld refinement from multiple powder data sets collected at different wavelengths for each sample using synchrotron X-rays. The powder data, collected in proximity to and far from the MnK absorption edge, allowed refinement of the fʹ anomalous scattering parameter for each data set along with other variables during the Rietveld structure analysis, and we could thus estimate the oxidation state of Mn from the valence chemical shift of the absorption edge. The results indicate that Mn in pumpellyite is distributed over both octahedral sites, and the refined anomalous scattering coefficients clearly show that the two Mn valence states are segregated: Mn 2+ is prevalent in the more symmetrical octahedral X site, and Mn 3+ is prevalent in the octahedral Y site. The present results confirm the relationship between site partitioning and oxidation state of transition elements in pumpellyite that was previously proposed to explain the distribution of Fe cations in the structure.

Energy-dispersive X-ray powder diffraction spectra have been collected for aragonite (CaCO 3 ) and dolomite [CaMg(CO 3 ) 2 ] at high pressure and high temperature, using synchrotron radiation and a cubic multi-anvil apparatus. Unit-cell volumes were measured up to 7 GPa and 1073 K along several isothermal paths. This study has confirmed previously determined but still debated values of the bulk incompressibilities at 298 K [65.4(5) GPa for aragonite and 90.7(7) GPa for dolomite] and thermal expansivitiesa t 1 bar [‹α› 298-1000 K = 6.7(5)10 -5 K -1 foraragoniteand [‹α› 298-1000 K = 4.1(5) 10 -5 K -1 fordolomite]. In addition, new equation-of-state parameters were measured: (∂K/∂T) p [ -0.013(2) and -0.025(4) GPa/K for aragonite and dolomite, respectivelyl and K′ 0 [2.7(7) and 2.3(5) for aragonite and dolomite, respectively]. We suggest that these equation-of-state parameters could be used in the calculation of high-temperature, high-pressure thermodynamic properties of these carbonates. At pressures exceeding 5 GPa, we observed that dolomite breaks down to aragonite and magnesite. We used the equation-of-state parameters measured in this study to calculate the position of the equilibrium curve: dolomite → aragonite + magnesite. This reaction could be important in the ultra-high-pressure metamorphism of carbonates. This decomposition reaction also provides a useful test of thermodynamic data sets of carbonates at high pressures and temperatures.

Defernite, Ca 6 (CO 3 ) 1.58 (Si 2 O 7 ) 0.21 (OH) 7 [Cl 0.50 (OH) 0.08 (H 2 O) 0.42 ], is a carbonate mineral [space group Pnam; at 100 K: a = 17.744(2), b = 22.601(4), c = 3.633(1) Å, Z = 4] with zeolitelike structural channels. The structural framework is formed by edge- and comer-sharing CaO 6-7 polyhedra, which are additionally linked by CO 3 groups aligned parallel to the (001) plane. This framework has channels parallel to the c axis, confined by eight-membered rings of CaO 6-7 polyhedra. The inner surface of the channels exposes H atoms of OH groups bonded to Ca. The framework carries a positive charge, which is balanced by small anions (mainly C1) in the structural channels. The crystal structure of defernite from the Kombat mine, Namibia, was refined using 3845 reflections to R = 2.6% at 100 K. The low but ubiquitous SiO 2 concentrations (1-5 wt%) observed in chemical analyses of this mineral were structurally located. Defernite has two symmetrically distinct CO 3 groups (C1, C2), which are repeated by a mirror plane parallel to (001). The C2 double layer of CO 3 groups is partially replaced by a disilicate Si 2 O 7 unit, where an additional O atom is accommodated between the two O triangles of the former CO 3 groups. This substitution is observed only for C2. In addition, all H atoms on the surface of the structural channels were located and refined. Cl in the center of the channel is bonded to six H within 2.8 Å. Defernite may form a solid solution between Ca 6 (CO 3 ) 2 (OH) 7 [Cl,OH] and Ca 6 (CO 3 ) 1.0 (Si 2 O 7 ) 0.5 (OH) 7 , in which the latter end-member is characterized by a neutral framework and structural channels without charged ions.

The crystal structure of the most K-rich natural pyroxene ever reported, a chromian diopside with 1.5 wt% K 2 O, has been refined (diffractometer data, filtered MoKα radiation, by least-squares using XTAL to R u = 3.2%) to examine the effect of K on the average structure. The crystal with structural formula Ca 0.80 K 0.073 Na 0.023 Mg 0.95 Fe 0.06 Cr 0.07 Al 0.02 Si 2 O 6 was found as an inclusion in a Koffiefontein diamond. The refined structure is typical of clinopyroxene on the diopside-enstatite join: Mg in M2 leads to distortion that is modeled by site splitting, with M2ʹ (Mg + Fe) displaced 0.33 Å from M2 (Ca + K + Na). Assignment of K to M2 is required to account for the electron density at that site. The average of eight M2-O distances (2.504 Å) is slightly larger than for diopside (2.498 Å). The effect of K on the average M2-O distance can be seen by calculating the average cation radius of atoms at M2 (+ M2ʹ) from the occupancy: 0.798·1.12 Å ( [8] Ca)+ 0.073 ·1.51 (K) + 0.023·1.18 (Na) + 0.036·0.92 (Fe) + 0.070·0.89 (Mg) = 1.127 ~≥ R (Ca). The large size of K is mostly offset by Mg + Fe in M2ʹ in the diopside-like structure; this sizedistortion balancing may facilitate K uptake in mantle clinopyroxene in K-rich environments. Large apparent thermal motion parameters of most atomic sites indicate sizable local distortions of the structure from substitution of K into M2.

In this paper, sample mass-absorption corrections for X-ray fluorescence trace element analysis are reexamined and a new approach is presented that is more accurate and more versatile than current methods. A method based on Compton scattering of a tube line is widely used because it is simple and does not require knowledge of the complete sample composition. The equivalent-wavelength method is sometimes used instead, especially if there is a major element absorption edge between the wavelengths of the Compton peak and the characteristic analyte radiation. Both methods suffer from difficulties in correcting for absorption edges and for enhancement by secondary fluorescence. In addition, the necessary assumption that the ratio of mass-absorption coefficients of any two elements is approximately independent of wavelength is surprisingly inaccurate. A series of examples demonstrates that in some cases large analytical errors may result. Methods based on the new approach completely and automatically correct for absorption edges and secondary fluorescence, without introducing such errors. In contrast to current methods, a complete major element analysis is not necessary. Thus, rapid determination of concentration ratios, such as Cr/Fe and V/Fe in oxide ores, Ba/Fe in Fe-rich hydrothermal deposits, and Sr/Ca in carbonates, is possible. In an experimental test, accuracy of better than 1% was demonstrated for analysis of V in Fe-rich samples. A proposed coefficient approximation was shown to give accuracy of 2% or better (excluding experimental errors) for Rb, Ni, Ba, and Cr over an extremely wide range of sample compositions.

Pressure-induced amorphous-phase transitions (amorphization) have been observed in framework silicates like quartz and feldspars. A similar crystalline to highly disordered solid phase transition is reported in this study for two zeolites, scolecite and mesolite. Pressure-induced amorphization is demonstrated using in situ X-ray diffraction and Raman spectroscopy. Before complete amorphization, progressive structural changes are observed below 10 GPa. The transition from the crystal to the amorphous phase is irreversible, and the pressure-quenched samples exhibit Raman spectra that bear strong resemblance to alumino silicate glasses obtained by quenching a liquid. In some cases, however, decompression from 15 GPa of fully amorphized samples is accompanied by a rim recrystallization of the sample and preservation of an amorphous phase with a structure different from that quenched in other experiments. Shear stresses seem to have important control of the transformation occurring during the decompression of the samples.

Thermodynamic properties of the natural calcium zeolites laumontite, leonhardite, dehydrated leonhardite (metaleonhardite), wairakite, and yugawaralite were studied by calorimetry in lead borate solvent at 975 K. Enthalpies of formation from the elements at 298 K are as follows: -7251.0 ± 8.5 kJ/mol for laumontite, CaAl 2 Si 4 O 12 · 4H 2 O; -7107.3 ± 5.6 kJ/mol for leonhardite, CaAl 2 Si 4 O 12 · 3.5H 2 O; - 5964.3 ± 5.1 kJ/mol for metaleonhardite, CaAl 2 Si 4 O 12 ; -6646.7 ± 6.3 kJ/mol for wairakite, CaAl 2 Si 4 O 12 · 2H 2 O; and -9051.3 ± 10.4 kJ/mol for yugawaralite, CaAl 2 Si 4 O 12 · 4H 2 O. The value for leonhardite is in good agreement with early values from acid calorimetry (Barany 1961) but not with revised values from Hemingway and Robie (1977). The enthalpy of dehydration of leonhardite is 140.2 ± 6.7 kJ/mol, and the loss of one mole of H 2 O is associated with an endothermic effect of about 40 kJ. Standard entropies, S 0 298 , of wairakite [400.7 J/(mol· K)] and yugawaralite [609.8 J/(mol·K)] were derived from our new enthalpy data combined with reversed P-Tphase equilibria (Liou 1971; Zeng and Liou 1982). The upper limit ofwairakite stability, the univariant curve for equilibrium of wairakite with anorthite, quartz, and fluid, was calculated from these values of enthalpy and entropy. Good agreement between thermodynamic calculations and reversed phase equilibria supports the reliability of the new thermodynamic data.

Leonhardite, a partially dehydrated laumontite, and its alkali variety, primary leonhardite, have been studied by high-temperature calorimetry. The enthalpies offormation from oxides and elements at 298 K are - 306.7 ±7.1 and -14214.6 ± 11.2 kJ/mol, respectively, forleonhardite, Ca 2 Al 4 Si 8 O 24 · 7H 2 O, and - 521.2 ± 10.5 and -14253.7 ± 13.5 kll mol, respectively, for primary leonhardite of composition Ca 1.3 Na 0.6 K 0.8 Al 4 Si 8 O 24 · 7H 2 O. The values for primary leonhardite are significantly more negative. New calorimetric data for sodium and potassium oxides were obtained on the basis of thermochemical cycles involving carbonates. The enthalpies of drop solution are -113.10 ± 0.83 kJ/mol for Na 2 O and -193.68 ± 1.10 kJ/mol for K 2 O, giving enthalpies of solution of -170.78 ± 0.90 kJ/mol for Na20 and -260.98 ± 1.20 kJ/mol for K 2 O. The effects of exchange cations (K,Na) on energetics and dehydration were studied using cation-exchanged samples. Alkali substitution decreases thermal stability (decomposition on heating in air) but increases thermodynamic stability with respect to the oxides and elements. Equilibrium relations between leonhardite and alkali-feldspar, calculated on the basis of these data, show that primary leonhardite can form only from geothermal solutions having rather high ratios of alkali ions to Ca.

The compositions of chlorite in divariant assemblages involving orthopyroxene + forsterite and spinel + corundum were reversed at a range of temperatures for P = 14 kbar and a range of pressures for T = 800 °c. These compositions, determined using a new calibration of the variation in the c unit-cell parameter with X chl , were then compared with those calculated using several simple activity-composition models for charge-balanced Tschermak substitutions in MASH chlorite. The discrepancies between the observations and calculations suggest that entropies of mixing are very low in chlorite solid solutions.

A combination of experimental and topological constraints indicates that critical equilibria play an important role in determining the pyroxene-liquid phase relations along the Mg 2 Si 2 O 6 -CaMgSi 2 O 6 (enstatite-diopside) join and in naturallherzolites. Topological constraints do not permit the three-pyroxene equilibrium, orthopyroxene + diopside = pigeonite, to intersect either of the two melting equilibria, pigeonite + diopside = liquid or pigeonite = orthopyroxene + liquid in the enstatite-diopside join. Thus, with increasing pressure, pigeonite cannot abruptly replace diopside on the solidus of simplified Iherzolites in this join. A supercritical low-Ca clinopyroxene, however, can replace diopside with increasing pressure along the solidus in the following manner: As the temperature of the solidus sweeps up the orthopyroxene + diopside two-phase field, the solidus crosses that region of the two-phase field where the diopside-rich limb of the two-phase field is inflected sharply toward enstatite by the proximal top of the metastable pigeonite + diopside solvus; once the solidus is at a higher temperature than the top of the metastable solvus, the diopside coexisting with orthopyroxene becomes supercritical and its Ca content drops rapidly but continuously. In the natural system and its synthetic analogs pigeonite may replace augite on the lherzolite solidus (orthopyroxene + augite + olivine + plagioclase + liquid) at constant pressure and decreasing Mgʹ [MgO/(MgO + FeO)] as the solidus liquid, which becomes less aluminous, intersects the liquid coexisting with orthopyroxene + pigeonite + augite, which becomes more aluminous with decreasing Mgʹ. However, with Mgʹ and the normative feldspar composition of the liquid nearly constant, the solidus liquid becomes sufficiently more aluminous with increasing pressure to prevent the three-pyroxene equilibrium from overtaking and intersecting the lherzolite solidus. Thus, in the natural system as in the enstatite-diopside join, pigeonite does not replace augite abruptly. Rather, the replacement of augite by low-Ca clinopyroxene with increasing pressure is relatively gradual and is influenced by the presence of a pseudoternary pigeonite + augite solvus. Experimental evidence suggests that the top of the solvus emerges from the local solidus at 13-15 kbar and forms a critical end point to the pigeonite + augite + olivine liquidus boundary. The emergence of the critical end point distorts both the remaining liquidus boundaries and the clinopyroxene solid-solution surface, such that the CaO content of the clinopyroxene (now supercritical) coexisting with orthopyroxene decreases more rapidly with further increases in pressure. The critical end point migrates to the orthopyroxenewollastonite join in just a few kilobars and completely eliminates the pigeonite + augite + olivine liquidus boundary curve by ~18 kbar.

Metapelitic rocks from a progressive metamorphic sequence in Dutchess County, New York, record evidence for reaction between garnet and fluids associated with quartz (± plagioclase) veins and for net transfer among garnet interiors, mineral inclusions, and fluid. Garnet-fluid reaction resulted in modification of preexisting garnet textures, compositions, and growth-zoning patterns. The textural and compositional record of garnet-fluid interaction varies with grade and with proximity of garnet to quartz veins. Garnet-zone rocks do not contain evidence for extensive reaction with vein-forming fluids. In upper staurolite- and kyanite-zone rocks, garnet that is crosscut by quartz veins contains a fluid-inclusion-filled region near the garnet-vein interface. This region truncates growth-zoning patterns and quartz inclusions. Fluid inclusions in garnet decrease in abundance away from quartz veins. In some rocks lacking quartz veins, fluid inclusions are concentrated in garnet interiors, and in particular near fractures, and are notably lacking around mineral inclusions in garnet. Inferences from textural and compositional features combined with thermobarometric results indicate that preexisting garnet grains were modified at T≈ 525-550 °C and >4 kbar by one or more reactions involving fluids. The fluid pathways included foliationparallel channels in the rock as well as microcracks in mineral grains.

Dehydrogenation experiments have been performed on natural pyrope megacrysts (Py 70 Alm 16 Gr 14 ) containing 22-112 ppm total H 2 O by weight (H/Si is from 0.00035 to 0.0018). The concentrations of OH in pyrope crystals were determined by Fourier-transform infrared spectroscopy. Two methods were used to obtain diffusivities. One method involved measurement of OH concentration profiles after a heating experiment. The diffusivity of the hydrous component was found to be proportional to the OH concentration along each profile, indicating that the diffusing species is not OH and is probably a minor or trace species with a concentration proportional to the square of the OH concentration (such as H 2 ). Hence, the diffusivities are referred to as the apparent diffusivities (D*, which equals D* 0 C/C 0 , where C 0 is the initial OH concentration and D* 0 is the D* when C = C 0 ). The other method involved measurement of the overall decrease of OH content across a wafer as a function of cumulative heating duration. From the overall decrease of OH content, the apparent bulk diffusion-out diffusivities (D̅* out ) were obtained. (D̅* in ) m, and D* 0 are different if D* depends on concentration.) The diffusivities from the two methods are consistent for a given pyrope crystal. However, the (D̅* out ) values in different pyrope crystals are roughly inversely proportional to the initial OH content in the crystal, opposite the concentration dependence of D* along a diffusion profile. This apparent paradox places strong constraints on the diffusion and incorporation mechanism of the hydrous component in pyrope and indicates a control on the concentration of the diffusing species by factors other than OH content (such as Fe 3+ /Fe 2+ ). Any proposed diffusion mechanism must be able to explain this apparent paradox (both the proportionality and the inverse proportionality). The (D̅* out ) values (in squared micrometers per second) in a crystal with 82-90 ppm initial total H 2 O can be described by In (D̅* out ) = [(28.20 ± 1.30) - (30580 ± 1450)]/T (where T is temperature in kelvins and errors are at the 20σ level). The (D̅* out ) values are greater by a factor of ~ 3 in a crystal with 22-23 ppm initial total H 2 O. The (D̅* out ) values can be determined by dividing the (D̅* out ) values by 0.347, and D* at each concentration can be determined from D* = D* 0 C/C 0 . The average activation energy for hydrous component diffusion in the two crystals is 253 ± 13 (2σ) kJ/mol. The diffusion-out diffusivities of the hydrous component are seven to eight orders of magnitude greater than the Fe-Mg interdiffusivities. They are so large that the OH content of a pyrope crystal can adjust to changing environmental conditions on a time scale of hours at temperatures as low as 800°C. Pyrope crystals from the mantle may dehydrogenate during ascent. Caution should be exercised in using OH content in natural pyrope crystals to infer conditions of the source region.

Hydrous partial melting experiments performed between 650 and 750 °C at 200 MPa (H 2 O) on synthetic metapelite compositions (quartz + albite + muscovite + biotite mineral mixtures) doped with Ba, Sr, Rb, and Cs yielded alkali feldspar crystals with a wide range of compositions in equilibrium at their rims with peraluminous melt. Measured partition coefficients for normally trace lithophile elements between feldspar and melt [D(M) Fsp/gl ,M = Ba, Sr, Rb, Cs] do not depend on either temperature or bulk composition of melt for the compositions studied. Values of D(Sr) Fsp/gl are between 10 and 14 and appear to be independent of the albite and orthoclase contents of the feldspar crystals. In contrast, values of D(Ba) Fsp/gl and D(Rb) Fsp/gl are strongly dependent on the orthoclase content of feldspar, relationships that can be expressed by the following linear equations: D(Ba) Fsp/gl = 0.07 + 0.25(orthoclase) and D(Rb) Fsp/gl = 0.03 + 0.01(orthoclase), where orthoclase is in mole percent. These equations reproduce the range of previously reported values for D(Ba) and D(Rb) determined on natural and synthetic samples. A single partition coefficient for Cs was also determined at D(CS) Fsp/gl = 0.13. These data can be used in conjunction with recently published partition coefficients for muscovite, biotite, and plagioclase feldspars (Blundy and Wood 1991; Icenhower and London 1995) to model quantitatively the trace element signatures of peraluminous magmas during anatexis and crystallization. Depending on bulk composition and pressuretemperature trajectories during episodes of partial melting, Ba and Sr (and to a lesser extent, Rb) are retained at the source while Cs and Li are enriched in the first melts. These results explain why peraluminous granitic magmas with S-type (sedimentary) sources carry a distinctive enrichment in Li and Cs.

A new member of the epidote group, androsite-(La), was found at Andros Island, Cyclades, Greece, as an accessory constituent of metamorphic Mn-rich rocks. Associated minerals are rhodochrosite, braunite, rhodonite, spessartine, and quartz. The mineral is brown-red in color, transparent with vitreous luster and very strong pleochroism. Lattice parameters are a = 8.896(1), b = 5.706(1), c = 10.083(1) Å, β= 113.88(1)°, V = 468.0(1) Å 3 , Z = 2, space group P2 1 /m. Crystal-structure refinement and chemical analyses yielded the ideal formula (Mn,Ca)REEAlMn 3+ Mn 2+ (SiO 4 )(Si 2 O 7 )O(OH). Another Mn-rich sample of the epidote group was found at Varenche Mn-mine, Valle di Saint Barthélemy, Western Alps, Italy; because of its chemical composition and crystallographic features, it can be considered an intermediate member of the piemontite-androsite-(La) series.

Hyttsjöite, circa Pb 18 Ba 2 Ca 5 Mn 2 2+ Fe 2 3+ Si 30 O 90 Cl · 6H 2 O, is a new mineral from the Långban mines, Filipstad district, Värmland, Sweden. It occurs sparingly in Mn-rich skarn with andradite, hedyphane, aegirine, rhodonite, melanotekite, calcite, quartz, potassium feldspar, pectolite, and barite. It is inferred to have formed below 300°C at 2-4 kbar from the breakdown of medium-temperature metamorphic assemblages in which hedyphane was the principal reactant. The name is from Hyttsjön, which is a lake situated west of the Långban mines. The average analysis is SiO 2 26.38, Al 2 O 3 0.02, Fe 2 O 3 2.77, CaO 4.35, MnO 2.00, PbO 58.13, BaO 4.65, Cl 0.65, H 2 O (calc) 1.58,O = Cl -0.15, total 100.40 wt%. Hyttsjöite grains are mostly 0.1-0.6 mm across, subequant to skeletal in outline, and colorless, and they have a prominent {001} cleavage. Optically, the mineral is uniaxial (-), ω = 1.845(4), ε ≈ 1.815.It is trigonal, space group R3̅, a = 9.865(2),c = 79A5(1)Å, v= 6695(3) Å 3 , and Z = 3; D calc = 5.10 g/cm 3 . The most intense X-ray (FeKa λ= 1.9373 Å) powder diffraction lines are as follows [d(Å)(I)(hkl)]: 13A(50)(006), 443(30)(0.0.18), 3.98(30)(027,1.1.12), 3.32(100)( 1.0.22,0.0.24,1.1.18), 3.11 (40)(217,128), 2.969(40)(2.1.10,0.2.19,1.0.25,1.2.11,0.0.27), and 2.671(80)(1.0.28,2.0.23,1.2.17). The crystal structure consists of composite SiO 4 + PbO n layers of two kinds (L1 and L2), which are supported by column segments parallel to c, the columns being composed of facesharing CaO 9 , FeO 6 , BaO 12 , and MnO 6 polyhedra. Each layer has two sheets of SiO 4 tetrahedra in pinwheel-like modules, joined together to form a puckered planar network filled out by PbO n groups. Layer L1 is continuous (Si 8 O 23 ), with intralayer Mn, whereas L2 is discontinuous (Si 7 O 22 ), with intralayer Fe and Ca. L1 and L2 are distinct from layers in other layered silicates, although they show some similarities to layers in minerals of the gyro lite group.

Nielson and Wilshire (1993) promised a review of geochemically based models for magma transport and metasomatism in the mantle. However, they focused on discrediting the chromatographic model ofNavon and Stolper (1987). Nielson and Wilshire concluded that the process operating during melt percolation in the mantle "resembles ion-exchange chromatography for H 2 O purification, rather than the model of chromatographic species separation proposed by Navon and Stolper (1987)." Our objective here is to show that Nielson and Wilshire (1993) did not offer any new theoretical or modeling advances. We demonstrate that the "water-purification model" (based on the finite-plate model of Helfferich 1962) is controlled by the same physics and mathematics as the frontal chromatography model discussed by Navon and Stolper (1987). Thus, the distinction made by Nielson and Wilshire between the two models has no basis. Moreover, we show that with appropriate boundary conditions, the model of Navon and Stolper (1987) produces an excellent fit to the data that Nielson et al. (1993) fitted using the "water-purification model." We emphasize that the model of Navon and Stolper can be used to model both small- and large-scale processes and that it allows the recovery of temporal information, which cannot be retrieved using the finite-plate model. The usefulness of these models for understanding the geochemistry of mantle rocks is a subject of active research and beyond the scope of this discussion.

Navon et al. (1996) demonstrated that the Navon and Stolper (1987) model can be formulated to reproduce a pattern of light-ion lithophile trace element (LIL) enrichments produced by a single, small-scale metasomatic process recorded in a composite xenolith from Dish Hill, California (Nielson et al. 1993). The Navon and Stolper model has failed repeatedly to reproduce the shape and lateral positions of LIL enrichment patterns for samples from peridotite massifs, which are of appropriate scale to test the assumption that LIL fractionation takes place in percolating melts over distances > 100 m. The model results also produce unreasonably long times for solidification of thin dikes, which imply untenable thermal conditions for lithospheric mantle. Using parameters drawn from sample compositions, Nielson et al. (1993) demonstrated, and the calculations of Navon et al. (1996) have shown again, that fractionated trace element patterns of a melt are imprinted upon relatively refractory peridotite matrix in zones closest to a melt source. The observed process sequentially extracts LIL into matrix, analogous to the ion-exchange chromatography of water-purification columns. We have never contended that this process is mathematically distinct from the percolation model of Navon and Stolper (1987), which assumes concentration ofLIL elements in melt. The choice of parameters defines the result, and one would notice a major difference in the taste of water from an ion-exchange column that traps target ions in matrix compared with one that concentrates those ions in the liquid. The difference between the models is in the selection of parameters and values: The model ofNavon and Stolper (1987) assumes the reaction mechanism, uses theoretical melt compositions, and contains as many as nine unmeasurable parameters. We used the simplified model calculation to avoid reliance on theoretical parameters and to test our assumptions about the process. When the compositions of actual samples are taken as end-members of mantle reactions, the successful results imply that a fractionation-bypercolation process is not applicable to lithospheric mantle. Repetition of the observed small-scale reaction in refractory peridotite must extend the zone of reactions and relative enrichment, centimeter by centimeter, as long as melt aliquots percolate beyond peridotite matrix that had previously reacted to equilibrium with the melt composition. This process satisfactorily explains the wide variations of LIL fractionation patterns over short distances that characterize mantle rocks in xenoliths and massifs, all of which contain complex systems of mafic intrusions with varied LIL fractionation patterns.

Assignment of the component bands in the infrared spectrum of tremolite containing A (Na,K) and T Al shows that the nontremolite components in the crystal are not randomly distributed throughout the structure but show extreme short-range order. Na and K at the A site are locally associated with Al at an adjacent T1 site.

A molecular dynamics study with analytical and ab initio interatomic potentials revealed the possibility of a new silica phase with Pnc2 structure, which is more stable than stishovite and its CaCl 2 -like modification at pressures above approximately 120 GPa over a wide range of temperature. Our analysis of simulated X-ray diffraction patterns of various hypothetical post-stishovite phases shows that this new phase should be considered in the treatment of high-pressure experiments involving silica. The new phase might be responsible for the seismic discontinuity at 2728 km depth and should be treated as potentially stable at the pressures and temperatures of the lowermost mantle.