Actinide diffusion in waste glasses

The diffusion of 233U and of 238Pu and 239Pu in phosphate glass, in different borosilicate glasses and in glass ceramics was investigated. Glasses with and without added simulated fission products were used. Some glasses also contained Gd2O3 as a neutron poison. The method of thin tracer layers and high resolution alpha spectroscopy were used to measure both depth and time dependence of diffusion. Diffusion rates were immeasurably small for T<3900C (diffusion coefficient, D<5x10-18cm2.s-1). Since the softening points of some glasses are as low as 5000C, only a limited temperature range was available for making accurate measurements. Above the softening points, the diffusion profiles became distorted due to plastic flow of the glass. Further complications arose due to nucleation of crystallites and subsequent crystallization of part of the glass for long annealing times and at high diffusion temperatures. The kinetics of the crystallization process, verified by optical microscopy and X-ray diffractometry, could be obtained from the time dependence of tracer penetration since differently crystallized glasses show different diffusion behaviour. In the crystallized glasses (glass ceramics), composite tracer penetration profiles were frequently observed. These could be unfolded to give a set of up to three Gaussians yielding different diffusion coefficients for different components of the glass ceramics. The effects of α-radiation damage and radiation enhancement of diffusion were also studied using curium and americium-containing glasses. In all cases, the measured D-values were small, indicating that, from the diffusion point of view, the glasses are a safe medium in which to store actinides. Observed plutonium gradients in leaching experiments are therefore stable until the gel-like surface layers get mechanically 'pealed off'. (author)