The influence of radiogenic 4He on cosmogenic 3He determinations in volcanic olivine and pyroxene

Abstract

Accurate determination of cosmogenic 3He concentrations in olivine and pyroxene phenocrysts requires knowledge of the amount of magmatic 3He also in the sample. The magmatic 3He component is commonly estimated by measuring the magmatic 3He/ 4He ratio and assuming that all 4He is magmatic. However, this approach yields incorrect results if 4He produced by U and Th decay is also present. Here we propose several strategies to account for the presence of radiogenic 4He. The optimal approach depends on whether the helium closure age ( T c) is similar to the exposure age ( T e) of the analyzed phenocrysts. (i) When T c = T e, which applies to uneroded lava flows, the ratio of cosmogenic 3He to radiogenic 4He is constant and the correction for radiogenic helium is independent of time. We provide a simple expression for a correction factor (R) that can be applied in this case. (ii) In the more common case that T c > T e, it is necessary to obtain an independent constraint on the closure age to estimate the radiogenic correction. In either case a quantitative estimate of the radiogenic 4He production rate is required. Because of the long stopping distance of α-particles, this production rate depends on the U and Th concentrations of both phenocryst and host, and also on phenocryst grain size. To illustrate the magnitude and uncertainty of the necessary corrections, we compiled U and Th measurements on phenocrysts and whole rock samples of basalts and andesites, supplemented by new measurements on Hawaiian basalts and Altiplano andesites. Our data and models suggest that some published cosmogenic 3He production rate determinations may have underestimated the true production rate by up to 5% because the presence of radiogenic 4He was not recognized. Similarly, a recent study presenting cosmogenic 3He derived erosion rates in > 4 Ma Hawaiian olivines probably overestimates true erosion rates by an order of magnitude or more.