Experimental and Theoretical Studies of Actinide and Lanthanide Ion Transport Across Supported Liquid Membranes

Abstract

In this work, we propose a new transport mechanism for metal ions relevant for used nuclear fuel separation processes by a supported liquid membrane (SLM). Two SLM extraction systems were investigated where the membrane was impregnated with either di-(2-ethylhexyl)phosphoric acid (HDEHP) or tributyl phosphate (TBP). A HDEHP impregnated membrane was used to extract neodymium (III), representative of a typical trivalent lanthanide. Cerium, which was oxidized by sodium bismuthate from trivalent to tetravalent state, was extracted by TBP. Oxidized cerium was used as a surrogate for oxidized americium to investigate the kinetics and possibility of americium and curium separation by membrane extraction. Both extraction systems were operated at varying nitric acid concentrations, and changes in the kinetics and extraction efficiency of metal ions were investigated. The proposed transport mechanism that was chosen for our studies was modified from the previous works by Danesi et al.[1,2] and Cussler et al.[3] The mechanism was selected due to the ability to accommodate and describe transport phenomena across a SLM when formation of extractant nano-channels in the membrane may exist. We were able to obtain acceptable fit of the models to our overall data trends although chemical and physical conditions must be well established and purity and homogeneity of the membrane are critical. A reverse transport of metal ions was observed when leaving the system for longer times which agrees with our model. The membrane was investigated for degradation and shown to be stable after contact with up to 7 M nitric acid for over 2000 minutes. Finally, we examined the possibility of partitioning americium from curium using a SLM impregnated by TBP. Separation of americium from curium was observed although not to a degree that was expected based on the Ce(IV) transport. Incomplete oxidation of Am(III) to Am(V) and reduction of Am(VI) on the membrane surface are possible causes for this observed discrepancy. Our model was, however, able to accurately predict Cm(III) transport through the membrane.