Garnet-biotite thermometry revisited: The effect of AlVI and Ti in biotite
Kleemann, Ulrich; Reinhardt, Jürgen
European Journal of Mineralogy Volume 6 Number 6 (1994), p. 925 - 942
published: Nov 30, 1994
manuscript accepted: Jul 5, 1994
manuscript received: Jan 11, 1993
ArtNo. ESP147050606019, Price: 29.00 €
Abstract Eleven published versions of the garnet-biotite thermometer have been tested by applying them to the experimental data of Ferry & Spear (1978) and Perchuk & Lavrent'eva (1983). Several calibrations show extreme deviations of up to about 260°C from the experimental temperatures. The statistical evaluation demonstrates that most calibrations published to date cannot be recommended for general use. On the basis of both experimental data sets, a new calibration was developed which uses the garnet activity model of Berman (1990) and a new activity model for biotite. This biotite activity model includes non-ideality of Mg,Fe-Al mixing (WH,MgAl - WH,FeAl = -77785 J/mol) and Mg,Fe-Ti mixing (WH,MgTi - WH,FeTi = 18138 J/mol), whereas Mg-Fe mixing was assumed to be ideal. In contrast to previous activity models, a strong temperature dependence of the non-ideality was found, with entropy interaction parameters WS,MgAl - WS,FeAl = -94.1 J/molK and WS,MgTi - WS,FeTi =11.7 J/molK. When applied to all experimental data, the new thermometer has the highest accuracy compared with the other 11 calibrations. Mean deviation and mean proportional deviation with reference to the experimental temperatures are -1°C/1.5 % for the Perchuk & Lavrenfeva (1983) data set and -6°C/4.5 % for the Ferry & Spear (1978) data set. Precision and accuracy has also been tested on natural rocks from several well-studied metamorphic sequences. The results show that the influence of octahedral Al and Ti in biotite on the temperature calculation is significant and cannot be ignored.