Automated calibration of laser spectrometer measurements of δ18O and δ2H values in water vapour using a Dew Point Generator

Niels C. Munksgaard, Alexander W. Cheesman, Andrew Gray-Spence, Lucas A. Cernusak, Michael I. Bird

    Research output: Contribution to journalArticle

    Abstract

    Rationale: Continuous measurement of stable O and H isotope compositions in water vapour requires automated calibration for remote field deployments. We developed a new low-cost device for calibration of both water vapour mole fraction and isotope composition. 

    Methods: We coupled a commercially available dew point generator (DPG) to a laser spectrometer and developed hardware for water and air handling along with software for automated operation and data processing. We characterised isotopic fractionation in the DPG, conducted a field test and assessed the influence of critical parameters on the performance of the device. 

    Results: An analysis time of 1 hour was sufficient to achieve memory-free analysis of two water vapour standards and the δ18O and δ2H values were found to be independent of water vapour concentration over a range of ≈20,000–33,000 ppm. The reproducibility of the standard vapours over a 10-day period was better than 0.14 ‰ and 0.75 ‰ for δ18O and δ2H values, respectively (1 σ, n = 11) prior to drift correction and calibration. The analytical accuracy was confirmed by the analysis of a third independent vapour standard. The DPG distillation process requires that isotope calibration takes account of DPG temperature, analysis time, injected water volume and air flow rate. 

    Conclusions: The automated calibration system provides high accuracy and precision and is a robust, cost-effective option for long-term field measurements of water vapour isotopes. The necessary modifications to the DPG are minor and easily reversible.

    Original languageEnglish
    Pages (from-to)1008-1014
    Number of pages7
    JournalRapid Communications in Mass Spectrometry
    Volume32
    Issue number12
    DOIs
    Publication statusPublished - 30 Jun 2018

      Fingerprint

    Cite this