Determination of pressure baseline corrections for clumped-isotope signals with complex peak shapes

Abstract. Pressure baseline corrections have been proposed to mitigate pressure-dependent background effects and reduce the apparent dependence of Δ₄₇ on δ₄₇ (non-linearity) observed in clumped-isotope studies of CO₂. In this work, we describe the determination of pressure baseline corrections for signals whose peak tops vary considerably across their width. Our study focuses on peaks with very small signal-to-baseline ratios (1.005 to 1.025) generated by the clumped isotopes ¹⁷O¹⁸O (linearly increasing peak top) and ¹⁸O¹⁸O (negatively curved peak top). The measurements were all performed in pure-oxygen gas using the compact, low-mass-resolution Elementar isoprime precisION Isotope Ratio Mass Spectrometer. We demonstrate that our corrections significantly reduce the influence of secondary electrons and that the adjusted clumped-isotope signals correctly increase with signal intensity. Furthermore, we extensively discuss correction procedures of varying complexity and explain why the best results were obtained by predicting multiple background values from the corresponding on-peak signals. Through this approach, we typically achieved standard deviations around 1 · 10^-9 (35/32), 0.2 ‰ (δ₃₅), 0.5 ‰ (Δ₃₅), 7 · 10^-9 (36/32), 0.1 ‰ (δ₃₆) and 0.1 ‰ (Δ₃₆) for at least 120 intervals (20 s integration). For the capital delta values, this corresponds to standard errors of the mean of less than 0.05 ‰, achieved with a total integration and analysis time of approximately 40 min and 6 h, respectively. We also show that the uncertainties of certain measurement parameters can be further reduced by optimising the measurement position (acceleration voltage) and applying additional drift corrections. For instance, for 35/32- and 36/32-related parameters we observed improvements of up to 1 order of magnitude and a factor of 7, respectively. Based on Monte Carlo simulations, we also show that the main uncertainties in our capital delta values are related to the on-peak signals, predicted backgrounds and the peak top curvature (only for Δ₃₆). Additionally, we present a brief study on the influence of pressure baseline corrections on major oxygen-isotope ratios and their delta values. While these corrections had an insignificant effect on their uncertainties, the absolute values of 33/32 and 34/32 changed markedly.