Oxygen isotope analysis of cellulose: An interlaboratory comparison

Oxygen isotope analysis of tree ring cellulose is hindered by the lack of a fast and reliable technique to measure delta(18)O at natural abundance levels. Most of the existing off-line techniques are not conducive to rapid generation of oxygen isotope data. Further, a comparison between oxygen isotope values of a cellulose sample measured using the different off-line techniques resulted in a discrepancy of 1.3 parts per thousand. This discrepancy is a result of inconsistencies inherent to these techniques and the lack of a reliable cellulose oxygen isotope standard. We addressed these problems by analyzing a number of cellulose samples by three different analytical procedures. These include two relatively well established off-line methods (pyrolysis in nickel tubes and decomposition using mercury(II) chloride) and a recently proposed online method based on pyrolysis in the presence of glassy carbon in an elemental analyzer and the subsequent isotope analysis of carbon monoxide. We were successful in improving the reproducibility of the on-line method (sigma(n-1) approximate to 0.2%0) to the extent that natural variations can be resolved and found a good agreement between the on-line and off-line methods, although the cellulose samples extracted from tree rings are subject to more scatter than the commercial cellulose samples. The on-line method may even allow determination of delta(13)C and delta(18)O of a cellulose sample in a single run (from 29/28 and 30/28 ratios of CO).