Synthesis of zircon-hafnon to determine oxygen isotope matrix effects in secondary ionization mass spectrometry

Abstract Secondary ionization mass spectrometry (SIMS) is a powerful tool for precise correlative actinide decay chain dating, trace element analysis, and stable isotope analysis of accessory minerals at unrivaled nanogram-scale sampling. Matrix-matched reference materials are a prerequisite for accurate quantification of isotopic compositions by SIMS. For rock-forming and accessory minerals showing partial or complete solid solution, elaborate correction schemes are required for SIMS isotope analysis. Natural zircon (ZrSiO4), often with a nearly stoichiometric endmember composition, has traditionally required less attention to matrix matching between reference materials and unknowns. However, with increasing analytical precision afforded by multi-collection SIMS instrumentation, it becomes important to experimentally verify this assumption and define its limitations. Here, we focus on Hf in zircon (Zrn), which is isomorphous with hafnon (Hfn), and the fourth most abundant element in natural zircon. Two endmembers in the Zrn-Hfn solid solution and three intermediate compositions were synthesized in a MoO2–LiMoO4 flux. Oxygen isotopic compositions of synthetic Zrn-Hfn crystals were determined at the milligram scale by laser fluorination isotope ratio mass spectrometry, and at lateral and depth resolutions of ∼15 μm and ∼1 μm, respectively, by SIMS. Despite a detected ∼1–3‰ isotopic heterogeneity in flux-grown Zrn-Hfn, a strong correlation between instrumental mass fractionation and the zirconium number Zr# % (atomic Zr/Zr + Hf × 100) was observed (Pearson correlation coefficient r = 0.9958), with ∼8.8‰ variation in δ18O across the compositional range. For most natural zircon, including common reference materials, the interpolated matrix effect is smaller than typical analytical uncertainties for individual SIMS spots (∼0.1‰). Only δ18O analysis of Hf-rich pegmatite zircon by SIMS requires significant (up to ∼3‰) matrix corrections. In such cases, the matrix effect on instrumental mass fractionation can be linearly interpolated between a common low-Hf zircon reference and the synthetic Hfn endmember to within ∼0.1–0.2‰ uncertainty.