Preparation of High-Quality Load-Preserved Fabric Clay Samples for Microstructural Characterizations: A Pragmatic Guide Featuring a 3-D-Printed Oedometer

摘要

High-quality samples are always critical in the meaningful microstructural characterizations of clay. A practical guide, together with a real sample tested as an example, is therefore proposed in this study, aiming to prepare high-quality load-preserved fabric clay samples subjected to 1-D consolidation for the pore size and fabric characterizations. In this guide, several issues often encountered in sample preparations that may induce measurement bias are tackled with feasible solutions proposed. A tailor-made oedometer is also invented and produced using a 3-D printing technique to achieve the goal. First, a homogenous and uniformkaolinite sample is prepared from a slurry state and then positioned in the 3-D-printed oedometer for 1-D consolidation tests. Then, together with the desired load, the whole 3-D-printed oedometer containing the consolidated kaolinite sample is submerged into liquid nitrogen to rapid freeze the sample to preserve the fabric associations not affected by the unloading effects. Afterward, to use the same sample for different tests to facilitate comparisons, the sample is cut in half while frozen. On one hand, this creates an observation plane along the center of the sample with the morphological information preserved for the subsequent scanning electron microscopy (SEM) analyses. On the other hand, each half of the sample individually undergoes the mercury intrusion porosimetry (MIP) and nitrogen adsorption (NA) analyses to obtain complementary information on the pore-size distribution. The samples must be dewatered by freeze drying before conducting these tests. During the freeze-drying process, the cut sample is protected by a 3- D-printed container, which has a reference mark to indicate the sample orientation. In each of the SEM images taken, the associated position and orientation are controlled, and the number of the images taken for analyses is maximized to enhance the statistical representation of the analyzed results.