Optimizing the Use of Carbonate Standards to Minimize Uncertainties in Clumped Isotope Data.

Clumped isotopes provide a theory‐based proxy to reconstruct formation temperatures of carbonates. With the introduction of the empirical transfer function (ETF) and several carbonate standards, interlaboratory comparison has become achievable. Due to the rare occurrence of the multiply substituted isotopologues, the analytical precision of these measurements is very low. In order to improve precision, we investigate the optimal strategy for the conversion into the absolute reference frame using a simulation approach. We find that changing the relative proportions of the standards to include more standards that are closer to the unknown target Δ47 value can greatly improve attainable precision. The inclusion of a hypothetical 4 °C standard results in only modest improvements in final temperature estimates for Earth surface temperature (0 and 40 °C) samples, indicating that the set of ETH‐1–3 standards is suitable for most applications. Full interpolation between two subsequent hypothetical standards, with Δ47 values of CO2 equilibrated at 1000 and 25 °C, results in modest improvements for samples with extreme Δ47 values. With a more optimal distribution of standards it is possible to measure more sample replicates before the uncertainty derived from the ETF becomes limiting. We provide suggestions for the optimal distribution of standards for all target sample Δ47 values and the R code to perform these simulations based on different laboratory settings. These optimizations can also be applied for ETFs using heated and equilibrated gases. We demonstrate numerically how optimizing the distribution and relative abundance of standards can increase measurement precision. Plain Language Summary: Carbonate clumped isotope measurements are a tool to reconstruct the temperature at which a carbonate has formed. Recent studies suggest that repeated measurements of carbonate standards allow for greater interlaboratory consistency and higher accuracy of measurements. In this study we demonstrate that the relative proportion of these standards should be changed to include more standards that are similar to the target clumped isotope composition of the unknown sample. We quantify how the composition of the standards used in a measurement session influences the measurement precision, based on simulations. Key Points: Measuring relatively more standards that are similar in Δ47 to the target sample improves precisionWith an optimal distribution of standards, the number of sample versus standard replicates should be approximately equal per session for smallest uncertaintiesAdding standards with extreme Δ47 values to the ETH‐1–3 set would lead to only minor error reduction at extreme Δ47 values [ABSTRACT FROM AUTHOR]