Experimental studies on amphiboles: a review with thermodynamic perspectives
Graham, Colin M.; Maresch, Walter V.; Welch, Mark D.; Pawley, Alison R.
European Journal of Mineralogy Volume 1 Number 4 (1989), p. 535 - 556
published: Aug 31, 1989
manuscript accepted: Mar 14, 1989
manuscript received: Aug 8, 1988
ArtNo. ESP147050104005, Price: 29.00 €
Abstract This paper addresses the geologically relevant question: are synthetic amphiboles useful analogues of their natural counterparts for petrological purposes? We address this question by considering the effects of starting material, bulk-composition and synthesis conditions upon the compositions and microstructural states of amphibole run products, and we observe that, as a consequence of incomplete synthesis yields and the difficulties in obtaining satisfactory chemical analyses, synthetic amphiboles may often be off-composition. Comparisons between natural and synthetic amphiboles are made by reviewing some well-documented case studies. Amphiboles can contain a variety of microstructural defects such as chain multiplicity faults and chain arrangement faults that can result in highly non-ideal structures. Some synthetic amphiboles can be highly defective as a result of nucleation and growth mechanisms. HRTEM study of anthophyllite-tirodite amphiboles, [Mg,Mn]7Si8O22 (OH)2, illustrates the point that major microstructural differences in chain disorder can exist between compositionally similar natural and synthetic amphiboles. This study has also shown the importance of the topotactic growth of amphibole upon a rapidly crystallised "precursor talc" in controlling amphibole microstructures. From other synthetic studies it is evident that microstructural state is very sensitive to composition as well as synthesis conditions; for example, synthetic alkali amphiboles such as glaucophanes and richterites grow in highly ordered micro-structural states. Attention is also drawn to the significant observation, realised through experiment, that theoretically univariant breakdown reactions of some important amphibole end-members are really divariant. This is illustrated by recent experimental studies on tremolite and synthetic "glaucophane". For example, tremolite breaks down divariantly thus: Tr = Trss + Diss + Qz + H2O. The limiting amphibole at the true univariant curve is a Mg-saturated tremolite: Trss = Diss + Enss + Qz + H2O. In a similar way, theoretically univariant breakdown reactions of glaucophane are actually divariant, involving ternary glaucophane-nyboite-Mg-katophorite solid solutions. The divariant nature of (potentially many) end-member breakdown reactions makes the derivation of well-constrained thermodynamic data difficult, and it may be necessary to resort to calorimetric methods, e.g. heat-capacity measurements, solutioncalorimetry. We also make some remarks concerning the potentially large errors in calculated amphibole equilibria arising from poorly constrained configurational entropies, taking pargasite as an example. The results of trial calculations on pargasite clearly demonstrate the need for good spectroscopic constraints on cation order-disorder schemes in amphiboles. A case is thus made for rigorous sample characterisation by HRTEM, crystal-structure refinements and spectroscopy, over and above the routine analytical methods currently used.