The J-CIREP was designed to simulate the degradation of refractories in alumina electrolysis cells.
The set-up used allows exposing a cylindrical refractory sample to a unidirectional flux of corrosive agents, namely, sodium vapor and fluorides. The sodium is generated by the reaction between the melt and liquid aluminum placed in a crucible on the sample (see Fig. 1). It then dissolves in the graphite crucible and diffuses toward the refractory sample. The melt percolates through the thin bottom section of the upper crucible to reach the sample. The rate of percolation can be controlled indirectly by carefully choosing the type of graphite used for the upper crucible (see Fig. 2)
With a simplified configuration, the sample is directly exposed to fluorides (see Fig. 3). In all these configurations, the refractory samples have the same geometry (7.6 cm diam. by 2.54 cm think) making it easier to compare the results.
The duration of a test can be from 24 to 96 hours or even more depending on the resistance of the tested material. The tests are performed under nitrogen (mainly to protect the crucibles) and preferably at 950°C.
After the test, the refractory materials is rated according to the extent of penetration observed, as well as to the nature of the chemical conversions that took place, based on XRF and XRD analyses.
Fig. 1: Schematics of the test configuration in which the refractory sample is almost only exposed to sodium vapor, by using a high density upper graphite crucible.
Fig. 2: Schematics of the test configuration in which the refractory sample is exposed both to sodium vapor and fluorides, by using a low density upper graphite crucible.
Fig. 3: Schematics of the test configuration in which the refractory sample is directly exposed to fluorides melt.
* For more information about this section, please consult the following reference:
PELLETIER, R., ALLAIRE, C., SILJAN, O.-J. and TABERREAUX, A, "Corrosion of Potlining Refractories: A Unified Approach", JOM, pp-18-22, August 2001.