PRESSURE-TEMPERATURE PATHS FROM GARNET ZONING: INTEGRATING GIBBS METHOD AND G-MINIMIZATION APPROACHES

Garnet growth zoning is commonly preserved in greenschist and amphibolite facies rocks. Extracting pressure-temperature (P-T) paths from the zoning may be done using differential thermodynamics (Gibbs method, program GIBBS), in which the initial conditions (P, T, the phases present, and their modes and compositions) are specified and growth is simulated by specifying changes in two variables that occur during garnet growth (e.g., ΔXalm and ΔXgrs), in accordance with Duhem’s theorem. Of the initial conditions that need to be specified, the only values that may be obtained through direct analysis are those of the garnet core composition. All others, including P and T, must be inferred. We inferred them from the bulk composition using a G-minimization approach (programs THERIAK and DOMINO). DOMINO plots equilibrium mineral assemblages, reactions, modes and compositions for specified intervals of P and T. The location of the garnet core composition on a DOMINO plot is unique and defines the initial P and T. All other initial values may be calculated by THERIAK using the initial P and T. Our experience reveals several characteristics: (1) The P-T locations of garnet cores do not lie on the zero garnet volume isopleths, but inside the respective fields of garnet growth. This may be due to the cut not intersecting the center, reaction overstepping, diachronous nucleation, or uncertainties in the calculations. (2) GIBBS simulations fractionate garnet whereas DOMINO does not, causing a mismatch of predicted compositions and modes along the P-T path that increases with increasing garnet grown (generally, negligible for garnet modes <10%). (3) Many of our garnets contain a hiatus in the P-T history that results from crossing the garnet-consuming reaction garnet + chlorite = staurolite + biotite along the path. The location of the hiatus may be recognized from an annular zone of elevated Y and depleted Ti that occurs at the beginning of the post-hiatus segment. The elevated Y may be due to garnet overgrowing matrix that became concentrated in Y that was released from the consumed rim, and the Ti depletion may be due to the partitioning of Ti into biotite during the garnet-consuming reaction. Combining Gibbs method and G-minimization approaches yields more tightly constrained P-T paths than is possible with either method alone.