Evaluation of sample pretreatment methods for boron concentration analysis

Abstract

Elemental boron (B) concentrations have been shown to be useful in reconstruction of paleosalinity conditions. Accurate measurement of the B content of a sample is critical for this purpose, but B is a light and relatively volatile element, substantial amounts of which can be lost through sample preparation protocols that fail to take this property of B into consideration. Here, we evaluate the effects of four commonly used sample preparation protocols on measured B concentrations. The Conventional Acid Method uses concentrated HNO₃ and HF, but the latter acid can cause volatilization of B. A second method, the modified Conventional Acid Method, uses the same acids but with the addition of H₃PO₄ to reduce B volatility. Two further methods eliminate HF entirely: the Ammonium Digestion Method makes use of NH₄F/NH₄HF₂, and the Alkali Fusion Method fuses the sample with Na₂O₂. We tested 9 samples and 4 standards using these four pretreatment methods. Our results show that the Alkali Fusion Method consistently has the highest B yields, with an average recovery rate of ∼98 %, whereas the other methods have average yields ranging from 72 % to 92 %. Owing to the potential for incomplete B recovery using these other methods (especially the Conventional Acid Method, which has the lowest recovery rate), salinity values based on those methods are likely to be underestimated. Given the dominant use of the Conventional Acid Method in older studies reporting B concentrations, paleosalinity estimates based on those studies will require re-evaluation. In the future, we recommend that all studies measuring B concentrations make use of the Alkali Fusion Method in order to achieve near-quantitative recovery, and to provide for maximum robustness and comparability of paleosalinity estimates among studies.