The recent discovery of heavy Si isotopic compositions in both high-Na Tonalite-Trondhjemite-Granodiorite (TTG) and High-K Granite-Monzonite-Syenite (GMS) suites of early continental crust requires that a notable seawater-derived silica-rich component had been added to their respective protoliths prior to melting. Here we use the Ge/Si ratio as a complementary tracer to delta Si-30 in order to delineate the exact role of modal quartz and silicified basalts from the Archean seafloor among the primary controls of the early appearance of felsic melts on Earth. We have approached the question by (1) specifying the Ge/Si signatures of various Archean and post-Archean rock types by compiling the Ge-SiO2 data stored within the GEOROC database; (2) coupling Ge/Si investigation to silicon isotopes on a large selection of silicified and unsilicified altered mafic and ultramafic greenstones, felsic volcanics, TTG and GMS granitoids and mineral separates from TTGs of the Barberton Greenstone Belt (BGB) of South Africa. The GEOROC compilation demonstrates that Archean TTGs and granites display much lower Ge/Si (1.15 +/- 0.10 and 1.13 +/- 0.11 mu mol/mol, respectively) than post-Archean adakites, granites, tonalites and granodiorites (with average Ge/Si in the range of 1.64 to 1.85 mu mol/mol). This result is corroborated by the Ge/Si ratios we report from BGB rocks that formed prior to Kaapvaal craton stabilisation, including 3.5 Ga Theespruit Formation felsic volcanic rocks, 3.5-3.2 Ga TTGs and 3.2-3.1 Ga GMSs, all of which exhibit low ratios (0.68 +/- 0.23; 0.92 +/- 0.17; 1.05 +/- 0.19 mu mol/mol, respectively). Based on their low TTG-like Ge/Si, precratonic GMSs are dismissed as being derived from melting of both TTG and metasedimentary sources. Instead, TTGs and GMSs originated from similar Ge-depleted sources. Low Ge/Si ratios, coupled to heavy Si isotopic signatures (-0.14 parts per thousand & nbsp;& nbsp;< delta Si-30 < +0.27 parts per thousand & nbsp;) are also a characteristic feature of mafic and ultramafic BGB greenstones subjected to low-temperature hydrothermal seafloor alteration, especially at the interface between silicified (0.2 < Ge/Si < 1.2 mu mol/mol) and unsilicified (1.8 < Ge/Si < 3.1 mu mol/mol) portions of the altered Archean seafloor. We infer that both Na-rich and K-rich Archean felsic melts are derived from a unique class of protoliths: Ge-depleted metabasalts containing a significant modal proportion of supracrustal quartz generated by the silicification of the Eo-Paleoarchean basaltic seafloors. The transition from Na-rich to K-rich felsic melts in the BGB is assumed to be connected to a gradual increase of potassium as a key element associated with seafloor silicification. In contrast, younger (3.07-2.69 Ga), post-cratonic BGB granites have higher Ge/Si (1.93 & PLUSMN;0.23 mu mol/mol) and were generated through the reworking of a TTG-like basement, by incongruent melting of biotite and hornblende (1.79 < Ge/Si < 2.97 mu mol/mol) leaving an oligoclase-rich (0.64 < Ge/Si < 0.72) residue. The Earth changed from a prevalent Ge-depleted felsic crust in early-middle Archean times to a widespread Ge-enriched post-Archean crust, which emphasizes the importance of late Archean changes in the Earth dynamics. In particular, our data suggest the likelihood of generating primitive felsic continents by rather shallow melting processes, without the need for inducing high pressures by subduction, as long as Archean ocean floors were silicified. |
