Kinetic control of staurolite–Al2SiO5 mineral assemblages: Implications for Barrovian and Buchan metamorphism

Abstract The distribution and textural features of staurolite–Al2SiO5 mineral assemblages do not agree with predictions of current equilibrium phase diagrams. In contrast to abundant examples of Barrovian staurolite–kyanite–sillimanite sequences and Buchan‐type staurolite–andalusite–sillimanite sequences, there are few examples of staurolite–sillimanite sequences with neither kyanite nor andalusite anywhere in the sequence, despite the wide (~2.5 kbar) pressure interval in which they are predicted. Textural features of staurolite–kyanite or staurolite–andalusite mineral assemblages commonly imply no reaction relationship between the two minerals, at odds with the predicted first development (in a prograde sense) of kyanite or andalusite at the expense of staurolite in current phase diagrams. In a number of prograde sequences, the incoming of staurolite and either kyanite, in Barrovian sequences, or andalusite, in Buchan‐type sequences, is coincident or nearly so, rather than kyanite or andalusite developing upgrade of a significant staurolite zone as predicted. The width of zones of coexisting staurolite and either kyanite, in Barrovian sequences, or andalusite, in Buchan‐type sequences, is much wider than predicted in equilibrium phase diagrams, and staurolite commonly persists upgrade until its demise in the sillimanite zone. We argue that disequilibrium processes provide the best explanation for these mismatches. We suggest that kyanite (or andalusite) may develop independently and approximately contemporaneously with staurolite by metastable chlorite‐consuming reactions that occur at lower P–T conditions than the thermodynamically predicted staurolite‐to‐kyanite/andalusite reaction, a process that involves only modest overstepping (<15°C) of the stable chlorite‐to‐staurolite reaction and which is favoured, in the case of kyanite, by advantageous nucleation kinetics. If so, the pressure difference between Barrovian kyanite‐bearing sequences and Buchan andalusite‐bearing sequences could be ~1 kbar or less, in better agreement with the natural record. The unusual width of coexistence of staurolite and Al2SiO5 minerals, in particular kyanite and andalusite, can be accounted for by a combination of lack of thermodynamic driving force for conversion of staurolite to kyanite or andalusite, sluggish dissolution of staurolite, and possibly the absence of a fluid phase to catalyse reaction. This study represents an example of how kinetic controls on metamorphic mineral assemblage development have to be considered in regional as well as contact metamorphism.