A thermodynamic model for Fe-Mg aluminous chlorite using data from phase equilibrium experiments and natural pelitic assemblages in the 100° to 600°C, 1 to 25 kb range

The purpose of this study is to derive a solid solution model for aluminous (Si < 3 a.p.f.u.) chlorites encountered in metapelites over a wide range of P-T conditions. A compilation of chlorite compositions in quartz-bearing rocks led us to propose a four-thermodynamic-component (Mg-amesite, clinochlore, daphnite, and Mg-sudoite) solid solution model that accounts for the Tschermak, Fe-Mg, and di/trioctahedral substitutions observed in nature. A new feature emerging from this compilation is the contrasting effect of temperature and pressure variations on the Al-IV and vacancy contents in chlorites. A 3-site mixing model with symmetric Margules parameters and ideal inter-site interaction has been adopted to model these compositional changes. In contrast to previous models, the relevant thermodynamic data (Mg-amesite and daphnite standard state properties as well as W-AlMg, W-AlFe, W square (Fe), W square (Mg), and W square (Al) on M1) are calibrated with independent sets of published experiments conducted in the MASH and FMASH systems (similar to 60 reversals) as well as about 200 natural data involving chlorite + quartz + (carpholite or chloritoid) assemblages. Moreover, the constraints span a wide range of pressure and temperature conditions (100 degrees -850 degrees C, 0.5-20 kb), so that no extrapolation outside the calibration range is needed for P-T thermobarometric purposes. The calculated thermodynamic data are compatible with the thermodynamic data of clinochlore from Berman (1988), Mg-sudoite and Mg carpholite data from Vidal and others (1992), Fe-chloritoid from Vidal and others (1994), and the chlorite-chloritoid Fe-Mg exchange thermometer of Vidal and others (1999). The chlorite solution model seems to be consistent also with the solid solution properties from Berman (1990) for garnet, Fuhrman and Lindsley (1988) for plagioclase, and Evans (1990) for epidote, although additional work is required to explain the large discrepancies observed between the temperatures obtained from empirical garnet-chlorite Fe-Mg exchange thermometers and the temperatures calculated in the present study. The use of several chlorite endmembers makes the estimation of paleo-pressure and -temperature conditions possible for high-variance parageneses (> 1) which is not possible when using only one chlorite endmember (classically clinochlore). In particular, reliable pressure estimates can be made for the common chlorite-quartz-carpholite or chloritoid or garnet bearing rocks devoid of aluminosilicates, whereas such estimates are impossible when using only one chlorite endmember. In the most favorable cases, temperature conditions can be estimated from the location of the temperature-dependent equilibrium 2 clinochlore + 3 Mg-sudoite = 4 Mg-amesite + 7 quartz + 4 H2O, that is from the composition of chlorite associated with quartz. Our chlorite solution model predicts that at fixed pressure and (XMg)(chlorite), the location of this equilibrium is shifted toward higher temperature when decreasing the Si, Al-VI, and vacancy contents and increasing the Al-IV content. This result is compatible with the classical empirical thermometers based on the Al-IV and vacancy contents in chlorite. However, the calculated effect of pressure is an increase of the Al-IV, Al-VI, and vacancy contents. This explains why the empirical chlorite thermometers (based on the Al contents in chlorite) derived from low-T samples cannot be used at high pressure conditions.