Dating agpaitic rocks: A multi-system (U/Pb, Sm/Nd, Rb/Sr and 40Ar/39Ar) isotopic study of layered nepheline syenites from the Ilímaussaq complex, Greenland

Abstract

The Ilímaussaq complex in southern Greenland is a shallow crustal composite intrusion comprising augite syenite, peralkaline granite and volumetrically dominant agpaitic nepheline syenites. Previous studies indicated a baddeleyite U-Pb age of 1160 ± 5 Ma for the augite syenite, the earliest intrusive unit of the complex. A similar crystallization age, within error, is inferred for the main sequence of agpaitic nepheline syenites. However, direct age determination of these units has been challenging because agpaitic rocks characteristically lack robust phases for in situ U-Pb dating (e.g. zircon/baddeleyite). An additional challenge is the pervasive subsolidus alteration, of which the isotopic effects are poorly constrained. Here we present new U-Pb, Sm-Nd and Rb-Sr isotopic data from whole rocks and mineral separates and a 40Ar/39Ar amphibole age of three co-genetic agpaitic nepheline syenites (kakortokite) from the lowermost exposed part of the complex. Using a multi-system geochronological approach for mineral separates and whole rocks, we explore the effects of late-stage alteration for each isotopic system. Assuming a closed-system evolution for the hydrothermal fluids (i.e. isotopically similar to the melts) and cooling within a relatively short time-frame (<0.8 Ma), we evaluate whether traditional mineral-whole rock isochron methods can provide useful age constraints for agpaitic rocks. We compare our data with those in the literature, corrected for the most recent decay constants.Single-crystal40Ar/39Ar step-heating experiments yield an amphibole plateau age of 1156.6 ± 1.4 Ma (MSWD = 1.5, external error ± 7.7 Ma), which we put forward as the most precise crystallization age for the agpaitic units to date. Kakortokite whole rock and mineral separates (amphibole, eudialyte, feldspar) yield a 206Pb-207Pb isochron age of 1159 ± 17 Ma (MSWD = 0.96) and a 235U-207Pb isochron age of 1168.5 ± 8.8 Ma (MSWD = 0.82). These are within error of the baddeleyite and zircon U-Pb ages from the augite syenite and alkali granite, as well as the new plateau age, if we take into account the external error of 7.7 Ma (i.e. accuracy). The 235U-207Pb age thus far provides the best non-single mineral age estimate for the agpaitic suite. Sm-Nd isotopes for the same whole rock and mineral separates yield an isochron age of 1156 ± 53 Ma (MSWD = 0.23) with εNdi = −0.8 ± 0.8, with significantly less scatter than Nd data for the rest of the complex. Rb-Sr isotopes yield errorchron ages that are either unrealistically young (3-point feldspar errorchron: 1106 ± 9 Ma, suggesting partial 87Sr loss), or old (WR, amphibole and eudialyte: 1237 ± 21 Ma, n = 9). The data demonstrate that the U-Pb and Sm-Nd systems are relatively insensitive to late-magmatic alteration and re-equilibration during cooling. In contrast, the Rb-Sr system records significant disturbance, reflecting the highly mobile nature of Rb and Sr in peralkaline systems. This warrants careful reconsideration of previously published Rb-Sr isochron data, and caution in interpreting Rb-Sr data for other peralkaline complexes. Initial isotopic compositions for the kakortokite support petrogenetic models that describe Ilímaussaq melt evolution towards strongly radiogenic Sr values at relatively constant εNdi, with progressive evolution from the early augite syenite to the most fractionated agpaitic melts. The melts experienced variable but minor degrees of lower crustal assimilation and preferential leaching of radiogenic Sr from the Proterozoic granitic country-rock.