Sensitivity of CSL Determination to Methods of Sample Homogenization Prior to Testing

A laboratory study was carried out on sandy silt gold tailings to investigate the effect of sample homogenization methods on the measured critical state line (CSL) in triaxial compression (TC) tests. Initially, 22 drained and undrained TC tests were performed on loose moist-tamped specimens prepared from the tailings that had been prepared using four different levels of homogenization: (a) without thorough mixing of the initial sample; (b) mixing during tamping of the specimens; (c) homogenization via sieving the material through 2.36 or 1.18 mm sieves; and (d) mixing the material as a slurry, air-drying it, and then sieving the material through 2.36 and 1.18 mm sieves. The results indicated fairly consistent CSL results within a void ratio of 0.03 for the first three sample homogenization methods employed, but showed a significant lowering of the CSL for the specimens homogenized by mixing as a slurry. An additional eight drained TC tests were carried out on dense moist-tamped and air-dried (AD) specimens to assess the effect of the lowering of the CSL on stress- and state-dilatancy behavior, indicating a lower state-dilatancy parameter, χtc, estimated from the AD tests than from the DMT tests.