TEM-EDS microanalysis: Comparison among the standardless, Cliff & Lorimer and absorption correction quantification methods

Available quantification methods for energy dispersive X-ray microanalysis in transmission electron microscopy, such as the standardless method (SLM), the Cliff-Lorimer approximation (CLA) and the absorption correction method (ACM), are compared. As expected, the CLA and ACM give superior results with respect to the SLM. As far as absorption can be considered negligible, CLA and ACM perform similarly. However, starting from mass-thickness of the order of 22 × 10−6 g/cm2, absorption become significant and the ACM gives better results. More accurate analyses can be obtained with the ACM if distinct kO/Si factors are determined for light and heavy minerals, respectively, placing a divide at 2.90 g/cm3. Caution must be used when k-factors are derived indirectly from minerals with very different structure/chemistry, suggesting that separate k-factors data sets are required for accurate EDS quantification, at least for the major and diverse broad classes of minerals. Element diffusion of monovalent cations and channelling effects may represent a complication, especially in very anisotropic minerals such as phyllosilicates, where these two phenomena may occur together.