The behaviour of Nd and Pb isotopes during 2.0 Ga migmatization in paragneisses of the Central Zone of the Limpopo Belt (South Africa and Botswana)

We have carried out major and trace element analyses and Nd-Pb isotope measurements on different components of two migmatite outcrops in the Limpopo Belt (Lose Quarry, Botswana, and Palala, South Africa) in order to assess the effect of migmatization during a high grade metamorphic event on Sm-Nd systematics and Nd model ages. The outcrops studied are located in the Mahalapye-Tshipise Straightening Zone at the south-western part of the Limpopo Central Zone close to its contact with the Kaapvaal Craton. Leucosomes of peraluminous granitic composition were formed by incongruent biotite melting in an Fe-rich metagreywacke (Lose Quarry) and in a metapelite (Palala), respectively. Monazite U-Pb and garnet Sm-Nd ages yield close to 2.0 Ga for the high grade event, confirming other recent work in the Central Zone. The chemical compositions of the main mineral phases (feldspar, biotite and garnet) indicate minimum melt formation at T > 750 °C and low pressure. Isocon plots and Pb isotope data indicate that the partial melts are locally derived. However, LIL trace element distribution between the leucosomes and paleosomes indicate the incomplete equilibration of the nascent melts with the residual feldspars and biotite in the paleosomes. Sm-Nd results on leucosomes and paleosomes of the metagreywacke migmatite show considerable scatter. When calculated back to 2.0 Ga, the paleosomes fall in a narrow range of ε_Nd = -4 and give T_DM model ages between 2.45 and 2.66 Ga. The ε_Nd values of leucosomes are on average four units lower, causing them to have apparent T_DM ages of 2.7-2.92 Ga, on average 250 Ma older than those of the paleosomes. The REE, Th, P₂O₅, and Zr budgets of paleosomes and leucosomes indicate dominance of the LREE by monazite and the HREE by apatite. Incorporation of proportionally more monazite than apatite into the melt is inferred as the cause for the Nd isotope discrepancy and hence the uncertainty of the average protolith mantle derivation age. In contrast, Pb-Pb results define an apparent age of 2025 ± 29 Ma, reflecting nearly full Pb isotope exchange. In the case of the Palala metapelitic migmatite, the ε_Nd values at 2.0 Ga of all but two migmatitic constituents cluster between -7 and -9.5, pointing out extensive Nd isotope equilibration during migmatization, whereas the Pb isotopes yield an errorchron date of 2311 ± 65 Ma, indicating incomplete Pb isotope exchange. Here, garnet dominates the HREE distribution whereas the LREE are controlled by feldspar and monazite. Considering that Nd isotope equilibration was nearly reached during migmatization, Nd model ages on the whole rocks represent a good approximation of those of the original protolith (3.0-3.2 Ga). In the regional context, the Pb isotope data from the Lose quarry locality combined with the Nd model age (uncertain, but within the range of 2.5-2.9) are reminiscent of the northern Marginal Zone of the Limpopo Belt and the southern Zimbabwe Craton, whereas the Nd and Pb isotope data from the Palala locality are typical for the Limpopo southern Marginal Zone and the Northern Kaapvaal Craton.