Tracking fluid sources during fluid-assisted metamorphism can be achieved by studying rare earth element (REE) behavior and isotope composition of accessory minerals. The hydrated zone within the Southern Marginal Zone (SMZ) of the Limpopo Belt in South Africa, contains metapelites (Bandelierkop formation) that experienced a ca. 2.7 Ga granulite-facies event during which the rocks underwent anatexis. The southern portion of the SMZ contains dispersed fragments of retrogressed metapelites with ubiquitous amphibolite facies assemblages. They include retrograde biotite, orthoamphibole, kyanite, and a 2nd generation of garnet that replaced orthopyroxene and partially replaced cordierite under pressure-temperature conditions of 660-600ºC and ≥ 0.6GPa. The hydrated metapelites contain graphite intergrown with the retrograde assemblages, indicating a mixed H2O-CO2 fluid and a rock-dominated system.
There are several hypotheses for the origin of the fluid involved in the retrogression process: 1) They derive from crystallizing in-situ melts that react with biogenic graphite in the metasediments; 2) They are mantle CO2-rich fluids that infiltrated the metapelites through deep shear zone systems after the granulite facies Neoarchean peak metamorphism; 3) Alternatively, ca. 2.1 Ga U-Pb rutile ages have been interpreted to argue for Paleoproterozoic retrogression, and that the fluids derived from Kaapvaal Craton sedimentary successions. Nonetheless, the origin of the fluids that triggered the retrogression hydration event is still unknown.
This study analyzed the chemical composition of 1st and 2nd garnet textures, monazite, apatite, and amphibole of the metapelites from the hydrated zone to constrain the fluid source and geochemical mobility of trace elements during fluid dissolution-precipitation reactions. Garnet REE patterns show two distinct populations described as Grt1 (Eu/Eu*=0.36) and Grt2 (Eu/Eu*=1.55). Monazite shows a relatively homogeneous LREE pattern for individual samples with a slight variation in HREE and negative Eu anomaly (Eu/Eu*=0.20-0.38). Apatite also displays distinct REE patterns, relatively enriched in REE (ΣREE = 1054-3009 ppm), depleted in REE (ΣREE = 141-159 ppm) or both. U-Pb monazite ages (ca. 2741 Ma-2707 Ma) suggest that the amphibolite-facies assemblages in the hydrated zone also experienced previous ca. 2.7 Ga granulite metamorphism recorded in the northern part of the SMZ. U-Pb apatite analyses yielded younger ages (ca. 2057 Ma-2047 Ma). Sm-Nd isochron of monazite, amphibole, and apatite from individual samples presented an initial 143Nd/144Nd ratio of 0.50894 ± 0.00006 (2s; n = 21; MSWD = 2.6), yielding a Sm-Nd isochron age of 2684 ± 51 Ma (2s). The Sm-Nd isochron indicates that monazite and apatite crystallized at ca. 2.7 Ga during the Neoarchean granulite facies event. Apatite experienced dissolution/reprecipitation during Neoarchean partial melting, as shown by the anatectic REE-enriched pattern. Amphibole grew during hydration of the SMZ. These mineral phases remained closed for Sm-Nd systematic, except for some apatite and amphibole that plotted slightly below the isochron.
We propose that the retrogressed zone of the SMZ experienced a Neoarchean peak granulite-facies followed rapidly by amphibolite-facies retrogression at ca. 2.7 Ga, in which an internally derived fluid interacted with the metapelites. This is supported by published Sm-Nd bulk rock compositions from samples of unretrogressed granulite facies and retrograde amphibolite facies rocks that follow the same evolution trend as our samples, ruling out the contribution of an externally derived fluid. A later lower-temperature Paleoproterozoic re-heating event was responsible for resetting the U-Pb apatite system by volume diffusion at ca. 2.05 Ga, consistent with the reported rutile U-Pb ages. Initial 87Sr/88Sr = 0.7130 ± 0.0014 (2s; n = 19; MSWD = 13) indicates a continental origin for the fluids that crystallized the apatite.