A refined protocol for δ18OPO4 analysis of conodont bioapatite

Abstract

The isotopic composition of phosphate-bound oxygen (δ18OPO4) in bioapatite (Ca10(PO4,CO3)6(OH,F,Cl)2) is increasingly used in paleoclimatology, paleoecology, and paleoceanography studies as a proxy for environmental conditions during mineral precipitation. However, a minimum total sample size (≥0.5 mg to >1 mg) required for requisite triplicate analysis can limit the spatial and temporal application of the proxy, in particular in studies using microfossils (e.g., conodonts) for δ18OPO4 analysis. Here we present a systematically evaluated extraction-conversion-analysis protocol for conodont δ18OPO4 of total sample mass as small as 0.3 mg while maintaining standard analytical precision (~0.3‰, 1σ). These results were produced by systematic testing of existing methodologies in combination with the vario PYRO cube high-temperature pyrolysis system. Our data indicate that bulk carbonate samples dissolved in 7% acetic acid buffered to a pH of 3.6 provide the highest conodont yields while minimizing bioapatite etching, thus avoiding potential oxygen isotope fractionation due to selective dissolution of conodont material. Isolating phosphate from the other components of dissolved bioapatite using a highly buffered silver amine solution produces larger silver phosphate crystals and greater phosphate yields compared to the standard silver amine protocol. Analysis of silver phosphate using a high-temperature pyrolysis system with reverse helium flushing and a thermodynamic carbon monoxide trap generates reproducible δ18O values (0.1 to ≤0.4‰, 1σ) on atypically small bioapatite samples (0.3 mg). Application of this refined technique to Middle Pennsylvanian age conodonts from Arrow Canyon, NV, demonstrates the potential of the conodont δ18OPO4 proxy for high temporal resolution (sub-104-yr) studies. Further, we propose a δ18OPO4 methodology to serve as a community standard that will permit more effective inter-dataset comparisons by limiting the potential influence of methodological variability on δ18OPO4 values.