Interlaboratory Variability of Radiocarbon Results Obtained from Blind AMS Analyses on Several Modern Carbon Samples

Abstract

Three samples of modern-day vegetation collected in 2009–2010 and a sample of bioethanol produced in 2010 were analyzed for radiocarbon by 5 different accelerator mass spectrometry (AMS) laboratories in a blind analysis study. The magnitude of any variability in the reported results for percent modern carbon (pMC) was observed. Results indicated that the interlaboratory repeatability on the samples of vegetation was generally very good, varying by no more than ~1 pMC for 2 of the 3 samples. Results for the bioethanol were less consistent, and varied by 5.5 pMC (ranging from 101.9 to 107.4 pMC). Variations in the 13C values used to correct for isotopic fractionation did not account for the variability observed in the pMC values for this sample. In view of the homogeneity of the bioethanol and its inherent simplicity in composition, this suggests that volatile liquid fuels may be more difficult to prepare for analysis without incurring significant sample processing errors. When viewing all of the results as a whole, the analytical errors (incorporating both instrumental and sample processing errors) appeared to be more random than systematic in nature. Because of analytical uncertainties in pMC measurements, as well as inherent local and regional variations in 14C activity levels known to occur in modern-day biomass, there is not a precise (accurate to 2 decimal places) correction factor for negating the bomb carbon effect that is applicable to all biofuels or other biobased products being analyzed in accordance with ASTM Method D6866. Therefore, a reasonable correction factor (currently set at 0.95) needs to be consistently applied in order to make comparisons of biobased content data from different laboratories more valid. Results from this study indicate that, for samples containing predominantly modern carbon, reporting results to the nearest 0.1 pMC is not warranted.