An assessment of the contamination contribution of various sample preparation procedures used at the Isotrace Radiocarbon Facility, University of Toronto, is described. Samples of geologic material, millions of years old, or samples derived therefrom, were tested because these would presumably contain only dead carbon. Results showed, however, that 14C contamination could be detected in several samples, complicating the contamination assessment. Best estimates of the contamination contribution from sample preparation were reported as: cracking: <0.17% modern, acetylene synthesis: <0.25% modern, combustion: <0.39% modern, and handling: <0.54% modern. These estimates were reported as upper limits because they likely represented 14C derived from two sources: sample preparation and the sample itself.

The levels and sources of the measurement background in an AMS 14C dating system have been studied in detail. The relative contributions to the total background from combustion, graphitization, storage, handling, and from the accelerator were determined by measuring the 14C concentrations in samples of anthracite coal ranging in size from 15 micrograms to 20mg. The results show that, for the present system, the uncertainty in the background is greater than that due to measurement precision alone for very old or for very small samples. While samples containing 100 micrograms of carbon can yield useful 14C dates throughout the Holocene, 200 to 500 micrograms are required for dating late Pleistocene materials. With the identification of the procedures that introduce contamination, the level and uncertainty of the total system background should both be reducible to the point that 100 micrograms of carbon would be sufficient for dating most materials.