Conductivity - Chemistry Relationship in Simulated Nuclear Waste Glass Melts
- PDF / 903,547 Bytes
- 6 Pages / 415.8 x 637.2 pts Page_size
- 9 Downloads / 203 Views
ABSTRACT Electrical conductivity of three simulated nuclear waste glass melts was measured over a wide range of temperatures, using a high-accuracy, calibration-free, coaxial-cylinders technique. Glass chemistries representing both low-activity and high-level wastes at Hanford and Savannah River sites were used. The temperature dependence of the conductivity of these glasses was presented. The conductivity values measured at 11 500 C were compared to the conductivity values predicted from first-order mixture model. INTRODUCTION Electric melting is a well-established mode of production of nuclear waste glass melts due to high thermal efficiency, ease of operation, yield of homogenous glass with the required properties, lower volatilization losses, and remote operation [1]. Electrical conductivity of glass melts is one of the most important processing parameters in the joule-heated melters used for vitrification of nuclear wastes. It should be in the range 10 - 100 S/m for successful melter operation. A lower conductivity at the melting temperature would require a higher voltage across the electrodes resulting in conduction within the melter refractory and could also cause melter start-up difficulties unless undesirably large electrical power system are supplied. A higher conductivity would result in current exceeding the recommended maximum density for the melter electrodes.
Two different electrical conductivity models for prediction of properties as a function of glass composition were reported in the literature. The approach taken by the first principles processproduct model by Jantzen [2] was based on glass structural considerations, expressed as a calculated non-bridging oxygen (NBO) term. Empirical first- and second-order mixture models were fit using a composition variation study (CVS) data (developed by the Pacific Northwest National Laboratory (PNNL) from 1989-1994)[3] to relate conductivity and other properties to glass composition. A model consisting of the Arrhenius equation with its coefficients (A and B) expanded in the forms of first-order mixture models accounted for 96 % of the variation in the CVS measured electrical conductivity data in the 950-1250'C range. For all CVS electrical conductivity testing, a probe with two platinum-l 0% rhodium blades was inserted into the glass to a known depth and the resistance of the glass between the blades was determined at discrete frequencies. For modeling, the resistance values at 1-kilohertz was used for consistency. Hrna and coworkers [4] developed mixture models for the logarithm of viscosity and that of electrical conductivity at 11501C. First-order models within the temperature range from 950 to 1250 0 C fitted the experimental data fairly well, accounting for roughly 98 and 96% of the variability in 'Department of Mechanical Engineering, Washington State University, Tri-Cities Campus, Richland, WA 99352 665 Mat. Res. Soc. Symp. Proc. Vol. 608 © 2000 Materials Research Society
the data, respectively. The authors reported the effects of glass components
Data Loading...
