Isotopic fractionation of zirconium during magmatic differentiation and the stable isotope composition of the silicate Earth

We thank the ERC under the European Community’s H2020 framework program/ERC grant agreement # 637503 (Pristine)) and for the UnivEarthS Labex program (no. ANR-10-LABX-0023 and ANR-11-IDEX-0005-02). Parts of this work were supported by IPGP multidisciplinary program PARI, and by Region île-de-France SESAME Grant (no. 12015908). High-precision double-spike Zr stable isotope measurements (expressed as δ94/90ZrIPGP-Zr, the permil deviation of the 94Zr/90Zr ratio from the IPGP-Zr standard) are presented for a range of ocean island basalts (OIB) and mid-ocean ridge basalts (MORB) to examine mass-dependent isotopic variations of zirconium in Earth. Ocean island basalts samples, spanning a range of radiogenic isotopic flavours (HIMU, EM) show a limited range in δ94/90ZrIPGP-Zr (0.046 ± 0.037 ‰; 2sd, n=13). Similarly, MORB samples with chondrite-normalized La/Sm of > 0.7 show a limited range in δ94/90ZrIPGP-Zr (0.053 ± 0.040 ‰; 2sd, n=8). In contrast, basaltic lavas from mantle sources that have undergone significant melt depletion, such as depleted normal MORB (N-MORB) show resolvable variations in δ94/90ZrIPGP-Zr, from -0.045 ± 0.018 to 0.074 ± 0.023 ‰. Highly evolved igneous differentiates (>65 wt% SiO2) from Hekla volcano in Iceland are isotopically heavier than less evolved igneous rocks, up to 0.53 ‰. These results suggest that both mantle melt depletion and extreme magmatic differentiation leads to resolvable mass-dependent Zr isotope fractionation. We find that this isotopic fractionation is most likely driven by incorporation of light isotopes of Zr within the 8-fold coordinated sites of zircons, driving residual melts, with a lower coordination chemistry, towards heavier values. Using a Rayleigh fractionation model, we suggest a αzircon-melt of 0.9995 based on the whole rock δ94/90ZrIPGP-Zr values of the samples from Hekla volcano (Iceland). Zirconium isotopic fractionation during melt-depletion of the mantle is less well-constrained, but may result from incongruent melting and incorporation of isotopically light Zr in the 8-fold coordinated M2 site of orthopyroxene. Based on these observations lavas originating from the effect of melt extraction from a depleted mantle source (N-MORB) or that underwent zircon saturation (SiO2 >65 wt%) are removed from the dataset to give an estimate of the primitive mantle Zr isotope composition of 0.048 ± 0.032 ‰; 2sd, n=48. These data show that major controls on Zr fractionation in the Earth result from partial melt extraction in the mantle and by zircon fractionation in differentiated melts. Conversely, fertile mantle is homogenous with respect to Zr isotopes. Zirconium mass-dependent fractionation effects can therefore be used to trace large-scale mantle melt depletion events and the effects of felsic crust formation. Publisher PDF