Defining an absolute reference frame for 'clumped' isotope studies of CO2

We present a revised approach for standardizing and reporting analyses of multiply substituted isotopologues of CO2 (i.e., 'clumped' isotopic species, especially the mass-47 isotopologues). Our approach standardizes such data to an absolute reference frame based on theoretical predictions of the abundances of multiply-substituted isotopologues in gaseous CO2 at thermodynamic equilibrium. This reference frame is preferred over an inter-laboratory calibration of carbonates because it enables all laboratories measuring mass 47 CO2 to use a common scale that is tied directly to theoretical predictions of clumping in CO2, regardless of the laboratory's primary research field (carbonate thermometry or CO2 biogeochemistry); it explicitly accounts for mass spectrometric artifacts rather than convolving (and potentially confusing) them with chemical fractionations associated with sample preparation; and it is based on a thermodynamic equilibrium that can be experimentally established in any suitably equipped laboratory using commonly available materials. By analyzing CO2 gases that have been subjected to established laboratory procedures known to promote isotopic equilibrium (i.e., heated gases and water-equilibrated CO2), and by reference to thermodynamic predictions of equilibrium isotopic distributions, it is possible to construct an empirical transfer function that is applicable to data with unknown clumped isotope signatures. This transfer function empirically accounts for the fragmentation and recombination reactions that occur in electron impact ionization sources and other mass spectrometric artifacts. We describe the protocol necessary to construct such a reference frame, the method for converting gases with unknown clumped isotope compositions to this reference frame, and suggest a protocol for ensuring that all reported isotopic compositions (e.g., Delta(47) values; Eiler and Schauble, 2004; Eiler, 2007) can be compared among different laboratories and instruments, independent of laboratory-specific analytical or methodological differences. We then discuss the use of intra-laboratory secondary reference frames (e. g., based on carbonate standards) that can be more easily used to track the evolution of each laboratory's empirical transfer function. Finally, we show inter-laboratory reproducibility on the order of +/- 0.010 (1 sigma) for four carbonate standards, and present revised paleotemperature scales that should be used to convert carbonate clumped isotope signatures to temperature when using the absolute reference frame described here. Even when using the reference frame, small discrepancies remain between two previously published synthetic carbonate calibrations. We discuss possible reasons for these discrepancies, and highlight the need for additional low temperature (< 15 degrees C) synthetic carbonate experiments. (C) 2011 Elsevier Ltd. All rights reserved.