To reconstruct the PT evolution of basement rocks, we first need to determine the paragenetic sequence in the high-grade schist and gneiss. After that, we can apply geothermometers and geobarometers. The results help place useful constraints on the PT history of these rocks, but also demonstrate that in most cases the minerals are not in equilibrium.
In the high-grade schist, we can identify an early post-S1 to S2 assemblage of quartz + biotite + garnet I + staurolite + plagioclase + Ti-Fe oxide (assemblage 1). Given the bulk composition of the schist, muscovite was almost certainly present, but is not now preserved. In the KFMASH system (K2O-FeO-MgO-Al2O3-SiO2-H2O), assemblage 1 corresponds to a divariant assemblage Grt + St + Bt + muscovite (Ms; Fig. 7), and it is therefore unlikely that an aluminosilicate was present in equilibrium with this assemblage (see Spear and Cheney, 1989).
The normal zoning pattern of garnet I suggests prograde metamorphism with increasing temperature and/or decreasing pressure, but the timing of this garnet growth is not clear from textural evidence. Resorption of garnet I may have occurred either as a result of the prograde univariant reaction Grt + Chl + Ms = St + Bt + Qtz + H2O, or during decompression, by the net transfer reaction Grt + Ms + H2O = St + Bt + Qtz, before staurolite had become unstable.
During D2, the stable assemblage in the high-grade schist is probably quartz + biotite + sillimanite + K-feldspar + ilmenite (assemblage 2). Staurolite became unstable when the conditions surpassed the univariant reaction St + Ms + Qtz = Grt + Bt + Sil + H2O, and muscovite breakdown occurred because of the dehydration melting reaction Ms + Pl + Qtz = Kfs + aluminosilicate (Als) + L (melt) (Fig. 7A). Textural evidence that support both reactions are the preservation of relict staurolite inside garnet II porphyroblasts, and the presence of K-feldspar + plagioclase + quartz intergrowths included in garnet II together with K-feldspar + plagioclase including blebs of andalusite. The latter two textures are interpreted as resulting from the later crystallization of residual melts, formed by the above-mentioned muscovite-dehydration melting reaction. Therefore, assemblage 2 results from the overstepping of two divariant assemblages in the KFMASH system, Grt + Bt + Sil + Ms and Grt + Bt + Sil + Kfs. Taken all together, this sequence indicates decompression, accompanied by constant or increasing T. Barometric estimates will provide additional evidences for decompression during D2.
Textural evidence suggests that the latest growth of garnet II occurs after D2, and therefore at lower P and/or higher T than the muscovite breakdown reaction. Garnet growth at this stage would involve staurolite breakdown under overstepped conditions, as suggested by the fact that some garnet II grains include corroded staurolite relics and include K-feldspar + plagioclase intergrowths. The only other garnet-forming reaction that could have been involved is the breakdown of biotite by the reaction Bt + Sil = Grt + Crd (Fig. 7A), which is ruled out by the absence of cordierite in the high-grade schist.
The final assemblage in these rocks is quartz + biotite + andalusite + K-feldspar + plagioclase + ilmenite (assemblage 3). This assemblage is interpreted to be formed by crystallization of residual melts in the andalusite stability field.
In the gneiss, the present assemblage is quartz + biotite + muscovite + sillimanite + andalusite + cordierite + plagioclase + K-feldspar + ilmenite ± garnet. This is clearly not an equilibrium assemblage, but the textural relationships do not allow an unequivocal paragenetic sequence to be established. Rather, the gneiss is best interpreted as a disequilibrium assemblage resulting from a series of strongly overstepped reactions that did not continue to completion.
The fibrolite seams probably reflect the result of the muscovite breakdown reaction (Ms + Qtz = L + Kfs + Sil or Ms + Pl + Qtz = L + Kfs + Als; Fig. 7B), leaving the assemblage Qtz + Bt + Sil + Kfs + Pl. The next step seems to have involved the partial breakdown of biotite, probably by a melting reaction that produced abundant Fe-Ti oxide (e.g., Bt + Sil = Grt + Spl + Mag + L; Clarke et al., 1989). Locally, garnet was probably produced by this reaction.
Textural observations support a final growth of cordierite, K-feldspar, and andalusite, probably representing the final crystallization of granitic melts. Large cordierite phenocrysts also occur with biotite and fibrolite inclusions. This textural evidence suggests a continuous growth of cordierite accompanying the melting reactions, and a final crystallization of melts in the andalusite stability field forming aggregates of cordierite, K-feldspar, and new biotite phenocrysts.
The observed disappearance of garnet I porphyroblasts, coupled with the appearance of cordierite in these rocks, suggests that cordierite could be partially formed by garnet-consuming reactions, such as Sil + Grt + Qtz + H2O = Crd (Fig. 7B). Continued growth of cordierite could be related to melting reactions, such as: Bt + Als + Qtz = L + Crd + Kfs (Fig. 7B).
The melt crystallization also produced new, Al-rich biotite, from the reaction L = Al-rich-Bt + Crd + Kfs + Qtz + H2O (Vielzeuf and Holloway, 1988; Fig. 7B). The coexistence of andalusite and a granitic melt can be explained by considering the expansion of the PT melt field because of the addition of B, F, and excess Al (see the summary in Johannes and Holtz, 1996). In view of the presence of tourmaline and monazite in the gneissic rocks, together with the Als + Pl-rich composition of leucosome veins, it is likely that the granite solidus in the rocks was displaced toward lower temperatures, within the andalusite stability field at low pressures (Fig. 7B).
The final stage in the evolution of the gneiss was growth of muscovite, perhaps because of the reaction between a residual hydrous fluid and the aluminous phases in the rock: Crd + Kfs + Als + H2O = Ms + Qtz and Kfs + Als + H2O = Ms + Qtz (Fig. 7B).
To determine the PT conditions achieved by the basement rocks we have applied different geothermometers and geobarometers. In order to avoid internal discrepancies in consistency between different calibrations, we have selected those using the same thermodynamic database.
Application of the garnet-Al-silicate-plagioclase (GASP) barometer (Hodges and Crowley, 1985) to plagioclase inclusions in garnet consistently gives slightly higher pressures in garnet I inner rims than in garnet II, which reflects in part the decreasing grossular component toward the rims. For an assumed temperature of 700°C, the calculated pressures are between 8.5 and 7.5 kbar, close to the kyanite-sillimanite transition (Fig. 7A). The garnet-plagioclase rims formed under lower pressures, but over a large pressure-range (3.5-5.5 kbar at 700°C). A similar pressure range is also obtained by using the garnet-plagioclase-biotite-quartz (GPBQ) barometer of Hoisch (1990). Pressure estimates using the GASP geobarometer for garnet I rims should be considered as maximum values, because of the lack of an Al-silicate in equilibrium with assemblage 1.
Garnet-biotite (GARB) thermometry (Hodges and Spear, 1982), applied to garnet-biotite rims in the matrix, gives us a wide variation in temperatures, from 650° to 850°C (at an estimated pressure of 4 kbar; Fig. 7A). The wide span in the calculated temperature (some incompatible with the stability field of the observed mineral assemblage) is further evidence of disequilibrium in these rocks (see García-Casco and Torres-Roldán, 1996). Another important factor to consider is that garnet homogenization by diffusion is enhanced by the high-grade conditions. This has occurred in some garnet rims and, therefore, these calculated temperatures should be used with caution. Many authors have also questioned the use of this geothermometer in high-grade rocks (e.g., Selverstone and Chamberlain, 1990; Spear and Florence, 1992). It is quite clear that the textural evidence (e.g., dissolution of garnet exposed different rim compositions, garnet is partially consumed by biotite in the matrix, diffusion processes, etc.), as well as the variability of the results, indicate that the GARB estimates are not reliable in these rocks. Therefore, the temperatures must be constrained by phase assemblages. Considering that the stable assemblage during D2 (assemblage 2: Sil + Kfs + Bt + Ilm) in the high-grade schist results from staurolite and muscovite dehydration melting reaction, we suggest that peak temperatures in these rocks during D2 were in the range of 650° to 700°C (Fig. 7A).
Keeping in mind these constraints, and taking the GARB estimates together with the results of the GASP and GPBQ barometers, we estimate approximate PT conditions of 8.5-7.5 kbar for garnet I rims and 3.5-5.5 kbar at 650°-700°C for the matrix S2 assemblage in the high-grade schist. Temperature estimates for garnet I rims could not be defined with the data presented in this paper, although phase relation modeling based on garnet I zoning patterns will help constrain the PT conditions and reaction history these rocks underwent at this stage (Soto and Platt, in press).
In spite of difficulties inherent to disequilibrium problems (spurious GARB results, overstepping reactions, etc.) we can propose an approximate and partial PT path for the high-grade schist during D2 and afterward. We have documented a limited drop in P during D2 (from maximum pressures of 8.5-7.5 to 3.5-5.5 kbar), achieving T between invariant points [I1] and [I2] at low P (Fig. 7A). The decompression PT path during D2 occurred probably under increasing T conditions, as documented by Soto and Platt (in press).
After D2, the rocks underwent a sharp temperature drop at low P (3-2 kbar). Metastable survival of early phases developed under higher pressure and temperature conditions (e.g., staurolite, garnet I without garnet II rims), together with the reconstruction of strongly overstepped reactions (e.g., staurolite preserved in andalusite), suggest that the cooling portion of the PT path proceeded probably at a very fast rate, which is also supported by age determinations in these rocks (see Kelley and Platt, Chap. 22, this volume; Monié et al., 1994).
Standard geothermometers and geobarometers (e.g., garnet-cordierite) in gneissic rocks always give spurious results. This result confirms our hypothesis that many of the mineral assemblages in the gneissic rocks are not in equilibrium. Any work on these rocks to determine the PT path should be done using available petrogenetic grids, and therefore the results should be considered qualitative. To account for melting reactions in metapelitic rocks, we have used the experimentally derived grid of Vielzeuf and Holloway (1988) in the KFMASH system (fluid-absent grid, XMg = 0.5) and other dehydration melting reactions derived by Vielzeuf and Clemens (1992; Fig. 7B).
Textural evidence suggests that melting occurred in these rocks because of the muscovite and biotite breakdown reactions. Muscovite dehydration melting reactions occur at pressures <4-5 kbar and temperatures >700°C (Fig. 7B). The intersection between the Ms-out reaction without plagioclase (reaction Ms + Qtz = Als + Kfs + L; Fig. 7B) and the granite melting curve defines an invariant point [I1] placed at ~725°C-5 kbar. Consequently, to form the granitic assemblage Qtz + Bt + Sil + Kfs + Pl, the gneissic rocks must surpass the band of reactions that define the Ms-out melting curve (dark gray area in Fig. 7B) at temperatures between the invariant point [I1] and those at the reactions involving melting of biotite to form pyroxene (Bt + Pl + Qtz = Opx + Kfs + L); that is, at T = 700°-750°C. P conditions for this assemblage are constrained by the presence of cordierite accompanying melting (i.e., P < 4-4.5 kbar; limit defined by the reaction Bt + Als + Qtz = L + Crd + Kfs).
All these observations together suggest a decompression PT path for the gneiss, at 700°-750°C, from at least 6 kbar to 3 kbar. Melting occurred along the decompression path by successive breakdown of muscovite and biotite. Cordierite was stable during melting, although its major growth occurred at low-P conditions, at the expense of biotite.
The decompression PT path was followed by a cooling path (P < 3 kbar) up to the andalusite stability field (T ~ 600°C). During this portion of the path, crystallization of residual melts occurred. These melts formed muscovite, new biotite (different in composition to restitic biotite), together with K-feldspar, plagioclase, cordierite, quartz, and andalusite. We interpret that a large overstepping of reactions occurred during the cooling path, resulting in complex textural relationships between these minerals. In particular, andalusite formed either as inclusions or including minerals of the granitic assemblage (Kfs + Pl + Crd). The contradictory textural relationships for andalusite in these rocks were previously interpreted as a result of heating under very low P conditions (Platt et al., 1996). Nevertheless, a careful textural interpretation of a more complete set of samples, together with the chemical characterization of the phases, allows us to reinterpret the PT path followed by these rocks.