When Did Plate Tectonics Begin on Planet Earth?
A trace element perspective on Archean crust formation and on the presence or absence of Archean subduction
Published:January 01, 2008
Stephen Foley, 2008. "A trace element perspective on Archean crust formation and on the presence or absence of Archean subduction", When Did Plate Tectonics Begin on Planet Earth?, Kent C. Condie, Victoria Pease
Download citation file:
The early continental crust is dominated by high-grade gneisses with the composition of sodic granites (the tonalite-trondhjemite-granodiorite or TTG suite) that date as far back as >3800 Ma. These are considered by many to be formed in subduction zones, and so have a critical role in the discussion about when plate tectonics may have begun. Trace elements can be used to learn about the identity of minerals in the source rocks during melting, but only indirectly to infer tectonic environments. The integrated results from experimental petrology and major and trace element geochemistry of the TTG suite indicate that most of them formed by melting of garnet amphibolites of broadly basaltic composition. This can explain low Nb/Ta coupled with high Zr/Sm, as well as low concentrations of HREEs. Melting of eclogite probably increased in importance in the Late Archean, as shown by an increase in Nb/Ta. Melting of garnet amphibolite can be achieved either in subduction zones at appropriate geotherms or in the lower reaches of thick basaltic crust. Water contents must be much higher than in the original basalts, indicating hydrothermal alteration at near-surface conditions. Thus, the thick crust scenario requires volcanic piling to deeply bury hydrothermally altered basalts, and also delamination of underlying thick cumulates. If subduction occurred in the Archean (implying the operation of plate tectonics), then with a slightly higher average mantle temperature, subduction geotherms would have been disproportionately hotter than today. However, there is no evidence for large volumes of TTG gneisses formed by melting of garnet-free amphibolites in the Early Archean, which constrains the average mantle temperature to be only marginally hotter than on modern Earth. In the Early Archean, melting of more magnesian volcanics and cumulates produced melts and prevented the production of voluminous continental crust with low-SiO2 >55 wt% SiO2. Given current trace element evidence, the most likely scenario for Archean tectonics is a slightly modified plate tectonics with only marginally hotter average geotherms, but substantially hotter subduction geotherms, enabling melting of garnet amphibolites in the Late Archean.