GARNET-BIOTITE THERMOMETRY — 40 YEARS LATER

The 1978 paper by Ferry and Spear “Experimental calibration of partitioning of Fe and Mg between biotite and garnet” (FS78) is arguably the most frequently applied geothermometer (>1400 citations). As such, a significant proportion of our understanding of crustal temperatures during orogenesis stem from its application. More recent studies typically forgo classical thermobarometry in preference to mineral assemblage diagrams (MADs). MADs were calculated for typical pelitic bulk compositions using the SPaC (2014) and Holland and Powell (1998) (HP98) datasets. Compositions of garnet and biotite from the MAD calculations were then used to calculate temperatures using the garnet-biotite thermometer (calibrations of Ferry and Spear, 1978, and Hodges and Spear, 1982). An unexpected result is that lines of KD between garnet and biotite [= (Fe/Mg)Grt/(Fe/Mg)Bt)] in MADS deviate significantly from the uniform, steep KD lines of the FS78 geothermometer. In the MADS, lines of KD may change abruptly crossing from one mineral assemblage field to another, and in some fields they may have a flat slope, implying 100% pressure dependence and 0% temperature dependence. For a typical low-Al metapelite in the KFMASH system, calculated differences (TGrt-Bt – TMAD) are -30 to +4, -68 to -24, and -90 to -9 C for the SPaC (ideal garnet), SPaC (Berman garnet model), and HP98 datasets, respectively. In the MnNCKFMASH system, the same calculations reveal differences of -69 to +120 and -62 to + 180 C for SPaC and HP98. Perhaps the most important cause for these discrepancies is that natural biotites contain octahedral Al, Ti, and Fe3+, often display octahedral vacancies, and Fe and Mg are ordered between the M2 and M1 sites. Inasmuch as octahedral Al in biotite depends on the coexisting phases and is often P sensitive, changes in the Al (or Ti) content will displace Mg from the M1 site preferentially over Fe and alter the bulk Fe/Mg. Other possible contributors are the imperfectly constrained Ca, Fe, Mg and Mn interaction terms in garnet. Fortunately, partitioning of Fe and Mg between M1 and M2 in biotite is relatively systematic with P and T (nearly insensitive to P). Therefore, Fe-Mg partitioning between garnet and the M2 site of biotite should be relatively systematic and could, potentially, provide a more robust geothermometer.