Determination of 18O by prompt nuclear reaction analysis: application for measurement of microsamples

A method is described for the routine determination of 18O concentrations in microsamples of biological fluids. The method utilizes the prompt nuclear reaction 18O(p, alpha o)15N, and 846-keV protons from a 3-MeV Van de Graaff Accelerator are focused on approximately 2,000-A-thick Ta2O5 targets prepared by anodic oxidation from 50-microliter samples of water distilled from blood or other biological fluids. The broad cross section of the resonance peak for this nuclear reaction (47 keV) ensures high yields, especially at small reaction angles, and the high-energy alpha particles produced by the reaction (4 MeV) are readily separated from scattered protons by the use of an aluminized Mylar foil of suitable thickness. Background levels of 18O (0.204 atom%) can be detected with run times of approximately 5-8 min, and the sensitivity of the method is of the order of 0.05 atom %. Experimental error due to sample preparation was found to be 1.7%, and counting errors were close to theoretical limits so that total error was of the order of 2.5%. Duplicate samples were analyzed by use of the 18O(p, alpha o)15N reaction at Lucas Heights, Australia, and the 18O(p,n)18F reaction by the method of Wood et al. (Anal. Chem. 47: 646-650, 1975) at the University of California, Los Angeles, and the agreement was excellent (y = 1.0123x - 0.0123, r = 0.991, P less than 0.001). The theoretical limitations and the general applicability of the method in biological studies designed to estimate the rate of metabolism of free-ranging animals are discussed.