Granite chemistry underwent a revolution in the 1980s. Distinctive patterns in major and trace-element compositions were recognized and categorized in such a way that the likely source materials and tectonic environments of granite magmas could be discriminated. Despite the relative ease and reliability with which granite provenance can now be ascertained, the causes of those patterns have not been identified except in the most general ways. Moreover, evolved granites and pegmatites accentuate those trace element signatures, but these are highly modified by fractional crystallization.
Our work on pegmatites has a complementary component in the origins of granite magmas at anatexis. By calibrating the stoichiometry of melting reactions and the partition coefficients for major and trace elements between minerals and coexisting partial melts, we can determine, precisely, how the trace-element signature of a granite melt is acquired. The same partitioning studies describe how the composition of a granitic liquid will change with crystal fractionation.
Figure 2. Fractionation of an incompatible trace element (purple sphere), from anatexis to saturation in melt. (A) The trace element signature of melt is controlled by partitioning between residual minerals and melt. (B) The trace element signature of melt is modified by fractional crystallization from the melt. (C) increasing concentration of the trace element finally achieves saturation of melt and precipitation of a mineral in which the trace element is an essential constituent, e.g,. Be in beryl.
At the pegmatitic stage, some elements that exist in the ppm range at the source become sufficiently concentrated to precipitate minerals of their own, e.g., as spodumene (Li), beryl (Be), pollucite, (Cs), cassiterite (Sn), and tantalite (Ta). Using melting relations and element partitioning, then, we can predict the concentration of a particular trace element at the beginning of granite magmatism and at the end pegmatitic stage (e.g., London et al., 1999). This allows us to model and answer an important and recurring question in the relations between granites and pegmatites – how much crystal fractionation must take place from anatexis to produce a pegmatite capable of precipitating spodumene, beryl, etc. (Fig. 2)? Most of our work has been on the elements of Groups IA (alkalis: Li, Na, K, Rb, and Cs) and IIA (alkaline earths: Be, Ca, Sr, and Ba – the behavior of Mg during crystal fractionation is sufficiently well understood).
We have also studied the behavior of added components to the granite system that may influence partitioning behavior. These include the important fluxes and ligands of boron (B), fluorine (F), and phosphorus (P). Like many of the economically important specialty metals, these ligands behave incompatibly in felsic crystal-melt systems and hence both are concentrated together in granite and pegmatite ore deposits. The association is so strong, e.g., between Sn and B, that many economic geologists presume that the specialty metal and exotic ligand form a stable complex in melt, and that this complex itself is responsible for incompatibility and concentration toward the ends stages of magmatism.