Development of delayed neutron technique as rapid and precise method for determination of uranium and thorium at trace levels in rocks and minerals, with applications to isotope geochronology

The method described was developed partly to provide information on the uranium and thorium content of samples which were to be dated by the uranium/lead, thorium/lead and lead/lead techniques. Since the method is non-destructive and involves no chemical preparation it has the prime advantage that the sample that has been analysed for uranium and thorium can also be analysed spectrometrically for lead; for the same reason blank errors are also obviated. The method is based on the detection of delayed neutrons emitted in the decay of the fission products of uranium and thorium. Fission is induced by exposing the samples to the intense neutron flux inside a nuclear reactor. The fast neutron component induces fission principally in 232Th and the slow neutron component in 235U, so that by making two determinations, one in the mixed flux, one with the slow neutrons screened out by a cadmium shield, both thorium and uranium may be determined. A single determination of both uranium and thorium occupies about six minutes; the lower limits of precise measurement are at present about 5 x 10-8 g/g for uranium and 2 x 10-6 g/g for thorium. A careful investigation has been made of the range of validity of the method when applied to the analysis of rocks and minerals, and it has been demonstrated that the adoption of proper procedures allows all but one of the possible interference effects to be reduced to negligible proportions. The remaining interference is due to the intense γ-ray flux emitted by the sample as a result of neutron activation of the sample matrix. The major advance made in the present work is the discovery that this interference, largely discounted by previous workers, can be very serious for geological materials, but that it can be reduced to an acceptable level by careful design of the neutron counter. A survey of the errors involved in the method and a discussion of the reproducibility of analysis is presented. (author)