Single laboratory comparison of MC-ICP-MS and N-TIMS boron isotope analyses in marine carbonates

Abstract

Measurements of boron isotopes (δ11B) in marine carbonates are an important tool for reconstructing past ocean acidity and atmospheric carbon dioxide levels, but are challenging to obtain due to low boron concentrations and the presence of only two stable boron isotopes. Previous studies have observed differences in absolute δ11B values of marine carbonates measured via negative thermal ionization mass spectrometry (N-TIMS) and multicollector inductively coupled mass spectrometry (MC-ICP-MS) that are relatively large compared to magnitudes of δ11B variability in paleoceanographic studies. Here we investigate the cause of these δ11B offsets and implications for paleo-pH reconstructions by performing N-TIMS and MC-ICP-MS measurements within the same laboratory on a suite of eighteen homogenized samples, including inorganic calcites grown under controlled pH conditions and Quaternary planktic foraminifera. Results show that δ11B values between both measurement techniques are highly correlated (R2>0.99), but calcite δ11B values are ~0.5 to 2.7‰ lower when measured by MC-ICP-MS compared to N-TIMS. The δ11B offset between techniques cannot be attributed to procedural blank contribution, but the magnitude of offset moderately correlates with indicators of sample chemical composition, including B/Ca, Mg/Ca and δ18O (R2=0.39 to 0.50). N-TIMS δ11B does not change with calcium addition to a biogenic carbonate sample, suggesting that varying calcium abundance between samples and standards is not the reason for N-TIMS “matrix effects”. Correlation between δ18O and the δ11B offset is particularly apparent for calcites (R2=0.71); the origin of this relationship is not clear, but is not likely due to 17O interference. Although identifying the origin of measurement offsets requires further study, we demonstrate that such offsets do not affect the validity of δ11B data: The sensitivity of δ11B in inorganic calcite to pH is the same when measured by either technique, and paleo-pH reconstructed from δ11B measurements in planktic foraminifera is equivalent within uncertainty when a technique-specific constant offset is applied. Our results strongly indicate that δ11B measurements from both measurement techniques (N-TIMS and MC-ICP-MS) can be employed for paleo-pH reconstructions with equal confidence.