The Tetragonal-to-Tetragonal Phase Transition in Gillespite-type Phases ­ Is There Elastic Softening at High Pressures?

Ronald Miletich (ronald@kristall.erdw.ethz.ch)

Lab. of Crystallography, ETHZ Zuerich, Sonneggstr. 5, CH-8092 Zuerich, Switzerland

Synthetic analogues of the barium-iron sheet silicate gillespite have been reported to undergo a tetragonal-to-tetragonal phase transition at high pressures (Miletich et al., 1998; 1999), a phase transition which is different from the "classic" tetragonal-to-orthorhombic one in gillespite (e.g. Hazen and Finger, 1983). For the Cr and Cu substituted endmembers the P4/ncc low-pressure polymorphs were found to transform to the P42₁2 high-pressure forms at ~2.3 and ~2.6 GPa.

Here we present high-precision compressibility data of both compounds that allow nature and characteristics of this phase transition to be determined. High-pressure investigations up to ~10 GPa were carried out in diamond-anvil cells by means of single-crystal diffraction using quartz for internal pressure calibration (Angel et al., 1997). The high-pressure X-ray diffraction studies were carried out on a synthetic Ba-Cr endmember sample and on natural effenbergerite crystals (= Ba-Cu endmember).

A small but significant hysteresis suggests that the character of the transition appears to be weakly first order. Both the P-V data and the pressure dependencies of the unit-cell axes show deviations from a typical compression behaviour thus yielding unexpected results for the equation-of state fits. The EoS fits for the low-pressure polymorphs (K_0,T= 55.9(6), 55.3(5) GPa, for K'=4) provide significantly too small (and even negative) values for K' (-2.8 < K' < 2.0) if data are fitted to a 3rd-order BM-EoS. The fits of the data of the high-pressure polymorphs result into values for K' of 8 to 10 and V₀ which are larger than the V₀ of the low-pressure polymorph. The anomalous compression behaviour is suggestive of elastic softening that appears to be responsible for the higher compressibility around the transition pressure, in particular between the transition pressure and ~6 GPa. The results of the single-crystal XRD measurements, which include high-pressure structure investigations, will be discussed in detail.

Angel RJ, Allan DR, Miletich R, Finger LW, J. Appl. Crystallogr, 30, 461-466, (1997).

Hazen RM, Finger LW, Amer. Mineral, 68, 595-603, (1983).

Miletich R, Sowerby JR, Angel RJ, EOS Transact, F834, (1998).

Miletich R, Friedrich A, Sowerby JR, Angel RJ, Eur. J. Min., Beih, 11, 156, (1999).

Experimental Investigation on Grain-Boundary Diffusion in Wollastonite Rims

Ralf Milke (rmilke@mail.gfz-potsdam.de) & Wilhelm Heinrich

GeoForschungsZentrum Potsdam, Telegrafenberg, 14473 Potsdam, Germany

Coronary structures in metamorphic rocks are produced by diffusion-limited growth of a product phase separating the reactant phases from each other. A widespread example are wollastonite rims between quartz and calcite (Joesten & Fisher, 1988; Heinrich, 1993). We have synthesized wollastonite rims according to the reaction calcite + quartz = wollastonite + CO₂ at 850 to 1100°C under nearly dry conditions. Runs were performed at 0.1 GPa in hydrothermal and internally heated gas pressure vessels, at 1 GPa in a piston-cylinder apparatus. From rim width measurements we derived values for grain-boundary diffusion coefficients by applying the relation X²(t) (alpha) µ.D^gb..(dRT)^-1, where X is rim width,  grain-boundary width, and d grain diameter. At 100 MPa starting from hot-pressed calcite with embedded quartz grains the initial wollastonite nucleation takes place on the quartz surface. The rim propagates both into the quartz grains and the CO₂-filled space around them. This is explained by counterdiffusion of Ca- and Si-species. In dry experiments at 1 GPa wollastonite initially nucleates on the calcite surface, independent whether quartz grains are embedded in calcite or vice versa. These wollastonite rims grow by replacement of calcite and thereby diffusion of the SiO₂-component. In experiments at 100 MPa the grain-boundary diffusion coefficients D^gb range from 1x10^-24 m³/s at 850°C to 1x10^-22 m³/s at 1090°C. At 1 GPa and 1100°C we measured 1.5x10^-21 m³/s (quartz in calcite) and 7x10^-21 m³/s (calcite in quartz). The larger values for diffusion through wollastonite rims nucleated on calcite grains result from their high porosity due to the migration of CO₂ from the calcite-wollastonite interface into the surrounding matrix. These grain-boundary diffusion coefficients are less than values derived from natural contact-metamorphic wollastonite rims (Joesten and Fisher, 1988) by 4 to 5 orders of magnitude. This suggests that the growth of the natural rims involved diffusion through an interconnected network of wetted grain-boundaries.

Joesten R & Fisher G, Geol. Soc. Am. Bull., 100, 714-732, (1988).

Heinrich W, Amer. Min., 78, 804-818, (1993).

An Investigation of Grain-Boundary Diffusion of Si, Mg, and O by SIMS Measurements on Isotopically Doped Enstatite Reaction Rims

Ralf Milke (rmilke@mail.gfz-potsdam.de), Wilhelm Heinrich & Michael Wiedenbeck

GeoForschungsZentrum Potsdam, Telegrafenberg, 14473 Potsdam, Germany

Grain-boundary diffusion is the rate-determining step for most exchange and net-transfer reactions over a wide range of metamorphic conditions. Grain-boundary diffusion coefficients are commonly measured by rim-growth experiments. A large experimental data set exists for the diffusion process in enstatite rims grown by the reaction forsterite + quartz = enstatite (Yund, 1997; Fisler et al., 1997). However, these investigations provide no definitive answer about the relative mobility of the diffusing components. In our experiments, synthetic forsterite (fo = 1) + quartz reacted to form enstatite rims at 1000°C and 1 GPa for 24 h under almost dry conditions. The quartz was isotopically enriched in ²⁹Si (94 mol-%) and ¹⁸O (62 mol-%). The £ 200 µm forsterite grains grew enstatite rims with thicknesses of circa 15 µm. Isotope concentration profiles were measured across the enstatite rims by SIMS. Profiles show that ²⁹Si is enriched in the outermost part of the enstatite rims. The ¹⁸O front advances the ²⁹Si front towards the forsterite. This demonstrates that Si is relatively immobile, which is in contrast to observations by Fisler et al. (1997) that suggest that SiO₂-diffusion is fast. It follows, that enstatite formation on forsterite does not occur by supply of the SiO₂-component, but by depletion of the MgO-component resulting from its outward diffusion from the forsterite grains to the enstatite-quartz interface. Thus, enstatite rim growth measured by Yund (1997) under conditions similar to our experiments probably refers to grain-boundary diffusion of the MgO-component. ¹⁸O shows diffusion profiles from high concentrations on the quartz side of the enstatite rim to non-enriched values towards the forsterite side, indicating that self-diffusion of ¹⁸O takes place in the opposite direction to the rate limiting diffusion of the MgO-component.

Fisler DK, Mackwell SJ & Petsch S, Phys. Chem. Miner., 24, 264-273, (1997).

Yund RA, Contrib. Min. Pet., 126, 224-236, (1997).