Comparison of isotope ratio measurement capabilities for CO2: Sample preparation and characterization by Isotope Ratio Infrared Spectroscopy

This paper will describe the characteristics and performance of a system to prepare up to ten 50 mL samples of pure CO₂ with on-demand ¹³C/¹²C ratios, together with an optimized calibration system for measurements by Isotope Ratio Infrared Spectroscopy (IRIS) that has allowed measurement of δ¹³C and δ¹⁸O values with 0.02 ‰ reproducibility (1 σ).

The needs for improved quality infrastructure and appropriate reference gases for CO₂ isotope ratio measurements has been a driver for recent research and development activities within the National Metrology Institutes, and the decision of the Gas Analysis Working Group of the CCQM to plan an international comparison (CCQM-P204) of capabilities of measurements of these quantities. The comparison will be coordinated by the BIPM, which has the mission of preparing the comparison samples, and the IAEA, who will assign their isotopic composition on reference scales. The BIPM has developed a preparation facility based on blending of different pure CO₂ sources of very different isotopic compositions, followed by cryogenic trapping and transfer to ten 50 mL cylinders. The target isotopic ratio ¹³C/¹²C can be adjusted by accurate flow measurements.

A Carousel sampling system with bracketing reference gas calibration and dilution system has been designed at the BIPM to allow rapid and accurate analysis of prepared gas mixtures by IRIS. A key feature of the calibration system is to maintain identical treatment of sample and reference gases allowing two-point calibration of up to 14 samples, and appropriate flushing protocols to remove any biases from memory effects of previously sampled gases. Measurements are performed by the IRIS analyzer at a mole fraction of nominally 700 μmol/mol CO₂ in air, by dilution of pure CO₂ gas controlled by individual low-flow mass flow controllers (0.07 ml/min), and with a feedback loop to control mole fractions to ensure that differences between references and sample gas mole faction stay below 2 μmol/mol. This level of control is necessary to prevent biases in measured isotope ratios, the magnitude of which has also been studied with a sensitivity study that is also reported.

The Carousel and IRIS measurements have been validated using pure CO₂ samples prepared with the gas blending facility, covering a range in delta values of -1 ‰ to -45 ‰ vs VPDB, and in all cases measurement reproducibility over several days of testing of 0.02‰ or better (1 σ) were achieved for both δ¹³C and δ¹⁸O, with negligible memory effects.

Samples produced and characterized with the facility will be distributed to institutes participting in the CCQM-P204 comparison exrecise, with measurements foreseen in the first quarter of 2020.