A standard material for liberation analyses and examination of the robustness of stereological correction procedures

In this work, an artificial standard material was developed and used to examine the robustness of various different stereological correction procedures. The four correction procedures that were examined were: large-sections correction, Hill's fast approximation, Barbery's correction and PARGEN correction.;There were three steps to this work: (1) refinement of the standard material--certain modifications were made to make the standard material (developed as an M.Eng. project) more versatile and easier to use. (2) sectioning and correction of computer-generated spheres--different liberation distributions of single-capped spheres were computer-generated and sectioned. The four correction procedures were used to correct the stereological bias in the sectioning data. The corrected liberation distributions were compared with the true liberation distributions. (3) sectioning and correction of standard material particles--the standard material was used to re-create the same distributions that were computer-generated. The particles were mounted, sectioned and the sectioning data measured with an electron microscope and image-analyzer. The data were corrected using the correction procedures and the corrected and true distributions were compared.;A two-phase (glass/lead borate) standard material was successfully developed. This standard material can be made to exhibit granular, layered or simple locking.;The correction of the sectioning of the computer-generated spheres and standard material particles yielded similar observations about the different correction procedures.;The large-sections correction provided a simple, uncorrupted correction. It performed better in the sphere cases than in the standard material cases.;Hill's fast approximation performed well except in the cases of narrow liberation distributions. The sectioning data of the standard material cases appeared to support the assumption in this correction that the locked section and locked particle distributions are identical.;Barbery's correction performed better in the standard material cases than in the sphere cases. The correction had problems in situations where an incomplete beta function could not be fitted to the true liberation distribution.;The PARGEN correction was able to provide a good estimate of the true amount of free material, but it had difficulty estimating the locked particle distribution.