Contents of Volume

Abstract

An isotopic investigation (Rb–Sr, U–Pb, Pb–Pb, and Sm–Nd) of the Archaean Union Mine pegmatite, Northern Province, South Africa, is reported. The intrusion of this pegmatite, devoid of zircon and monazite, has been dated at 2912±2.6 Ma by Pb–Pb step leaching of magmatic garnet. This age is supported by an identical two garnet–apatite Sm–Nd isochron of 2917±27 Ma. A Rb–Sr K-feldspar–garnet-albite isochron of 2023±11 Ma corresponds to a later (hydro)thermal overprint. Isotope resetting during this overprint was incomplete and revealed that various parent–daughter isotope pairs within minerals of the Union Mine pegmatite behaved in exceedingly different manners under the same physico-chemical conditions.Sr isotope resetting was incomplete and resulted in open system behaviour of the pegmatite. K-feldspar almost quantitatively lost its (radiogenic) Sr, which had accumulated between the time of intrusion and the overprint. Sr loss from muscovite was variable and incomplete as indicated by a wide range in ages from 2088±74 Ma to 2744±13 Ma. Only insignificant portions of this Sr were incorporated into adjacent albite (as revealed by a micro-drilled Rb–Sr profile through a muscovite–albite grain boundary). The 87Sr/86Sr initial ratio is 0.8301±11 for the Proterozoic isochron, which is far less radiogenic than the calculated value of 2.24 for a closed system. It is therefore concluded that Sr was lost from the system and that the Rb/Sr ratio of the whole-rock was changed. Apatite, on the other hand, seems to have behaved as a closed system and retains a low initial Sr ratio of 0.7125.U–Pb analyses of apatite reveal that this mineral has behaved as an almost closed system for Pb diffusion as well, but was open for U (significant loss) during the overprint. This has resulted in high reverse discordance of apatite U–Pb analysis and in a geologically meaningless secondary Pb–Pb isochron.The data demonstrate that both preferential loss of daughter (Sr) over parent (Rb), and parent (U) over daughter (Pb) isotopes can occur in some minerals of one and the same rock. This process altered the parent/daughter ratios to such an extent that, in the case of apatite, a secondary Pb–Pb isochron with an erroneous age of 4119±66 Ma was established. It appears likely that preferential loss of daughter or parent isotopes is the reason for `spurious' secondary whole-rock isochrons reported and reviewed by Moorbath and Taylor (1986)and Moorbath et al. (1986). Such losses, far exceeding predictions based on volume diffusion experiments, may occur by means of fast-transport diffusion, which in part may depend on the micro-textures of the host minerals.Geochemical studies of early Archaean rocks require substantial back-correction for radiogenic decay, and the example of the Union Mine pegmatite shows that changes of parent/daughter isotope ratios may occur long after rock formation. However, such changes may go unnoticed in high precision geochronology (i.e. garnet dating in this study), but will lead to erroneous correction for radiogenic decay and hence to erroneous conclusions regarding the origin of the studied samples.