Liquidus and sub-liquidus phase equilibria for an Archaean tonalite: matching experimental data to models of TTG genesis

Experiments at 1050 °C and 2.0 GPa on two Eoarchaean (4.3 Ga) greenstones from northern Quebec in Canada (Adam et al., 2012; Geology; 40; 363-366) produced partial-melts that were compositionally similar to an associated but younger (3.7 Ga) tonalite (PC-103) of the TTG series, leaving a residual mineral assemblage of garnet and clinopyroxene. During this present investigation, a complementary experimental study of PC-103 produced matching liquidus phase equilibria; thus with 5 wt. % of dissolved H2O,the tonalite's liquidus is saturated with both garnet and clinopyroxene at ~1050 °C and 2.0 GPa. These results provide support for the popular belief that TTGs were produced by partial-melting of an earlier basaltic protocrust. In spite of this, alternative interpretations remain tenable, particularly when the broader experimental data base is considered. Of 11 published experimental studies of gabbro/amphibolite/eclogite melting (involving 17 different starting compositions and over 126 individual experiments) almost none have succeeded in more than partially matching the compositional characteristics of natural TTGs. Indeed, when compared to the compositional spread of the experimental melts, the data for natural TTGs define a relatively narrow and consistent pattern of variation. Thus, either very specific physical conditions and protolith compositions were involved, or TTGs were produced by some other mechanism or combination of processes. At a pressure less than about 1.0 GPa, the liquidus of PC-103 is saturated with both plagioclase and orthopyroxene. Thus, it is possible that its composition was controlled by fractionation along the plagioclase-orthopyroxene cotectic at mid to lower crustal conditions. This is confirmed by the compositions of melts produced along the plagioclase-orthopyroxene cotectic at lower temperatures, since these closely match the compositional trends of natural TTGs. In this case, it must be assumed that the distinctive incompatible element characteristics of PC-103 (and thus most other TTGs) were inherited from more mafic parent magmas, and that they were not a direct result of fractionation during protocrust melting.