A large scale comparative study of stable carbon isotope ratios determined using on-line combustion and low-temperature pyrolysis techniques

The stable isotope ratios of carbon and oxygen are normally measured separately, by on-line combustion- and pyrolysis-IRMS respectively. Measuring both isotopes simultaneously, using pyrolysis, would greatly improve efficiency. Results are presented from a large scale comparative study (> 2500 samples) of carbon isotopic ratios (δ13C) obtained by both on-line combustion and by pyrolysis at 1090 °C. The data resulting from the two methods are very strongly correlated (r = 0.91, p < 0.001), the mean isotopic values are statistically indistinguishable and more than 45% of the paired values of individual tree-rings fall within ± 0.2‰. The pyrolysis results, however, exhibit significantly lower variance, probably as a result of exchange with carbon deposited in the reaction tube. The δ13C results obtained by low-temperature pyrolysis are therefore less accurate than those obtained by combustion; low values being slightly overestimated and high values slightly underestimated. However, for most palaeoecological applications, where adequate replication is more critical than absolute precision, or where the amount of sample material is limited, this disadvantage is far outweighed by the efficiency savings. Differences in variance associated with the two techniques would not affect any environmental reconstruction based on regression, as the regression model would rescale the data accordingly. Simple precautions are proposed that will reduce the impact of this observed reduction in variance.