GLAUCOPHANE SYNTHESIS: THE ROLE OF WATER

The inability to define the P-T stability field of glaucophane, a key index mineral of the blueschist facies, has motivated experimental work on this amphibole for many years. Pioneering work in the early 1960's by W. Gary Ernst on amphibole synthesis attempts with end-member glaucophane (Na2Mg3Al2Si8O22(OH)2), including the (then) recently-developed piston-cylinder press, laid important, if not perplexing, groundwork for subsequent research. Ernst used the decrease in unit-cell volumes with increasing pressure to propose a transition boundary between two polymorphs of glaucophane. Subsequent work over the next 40 years concluded that the amphiboles that are formed at 0.8 - 3.2 GPa and 600-800° C are not ideal glaucophane and that there is no polymorphic transition. The largest obstacle to forming glaucophane appears to be composition/kinetic related. This study specifically considers the role of water.

Reconnaissance experiments over the range of 1.2 to 5.0 GPa, 650-775°C, for 2-5 days using a stoichiometric mixture of oxides sealed in Pt capsules with 0-30 wt% H2O were done at U. Alberta and Binghamton U.. The results agreed with earlier studies; glaucophanic amphibole appears with other phases between ~1.5 and 3.0 GPa. The highest amphibole yields (>90 wt%) were made with low amounts (<15 wt%) of water. A detailed series of experiments were done at 2.5 GPa and 750°C by sealing the starting mixture with 0-15 wt% H2O. Runs with greater than ~7 wt% H2O yielded charges that were soft and had the assemblage amph+smect+jadeite+qtz, while runs made with less than ~7 wt% H2O had hard charges consisting of amph+jadeite+talc. Microprobe analysis showed an increase in the Mg and decrease in the Si content of amphiboles, while unit-cell dimensions increased slightly, with increasing water content. The most glaucophane-rich amphiboles (Na1.91(8)Mg3.12(13)Al1.97(12)Si7.98(5)O22(OH)2) required 2-3 successive treatments (at 2.5 GPa, 750 °C) with 3-5 wt% H2O and intermittent grinding to eliminate jadeite and nearly all talc. Combining the composition and cell-volume data of this study with that of Pawley (1992) for nyböitic amphibole and the known volumes of Mg-cumm (Hirschmann et al. 1994) and Mg-richt (Cámara, pers. comm.), one obtains cell volumes of 875.5 and 864.0 ų for ideal nyböite and glaucophane, respectively.