Use of scanning electron microscopy and microanalysis to determine chloride content of concrete and raw materials.

PDFTech ReportBDK75 977-15Chloride contentConcreteCementElectron microscopyElectron microscopesSpectroscopyMicroanalysisFloridaFlorida. Dept. of TransportationDempere, Luisa MeliaWillenberg, BradleyDeist, RichardUniversity of Florida. College of EngineeringNTL-HIGHWAY/ROAD TRANSPORTATION-HIGHWAY/ROAD TRANSPORTATIONUS Transportation CollectionStandard sample sets of cement and mortar formulations with known levels of Cl as well as concrete samples subject to Cl diffusion were all prepared for and analyzed with scanning electron microscopy (SEM) and electron microprobe (EPMA). Using x-ray spectroscopic data generated with these instruments large scale (cm\uc2\ub2) elemental distribution maps were constructed for the standard samples sets and used to calibrate each instrument for future quantitative analysis of cementitious samples. Quantitative distribution maps using both the SEM and EPMA instruments were constructed for all concrete samples subject to Cl diffusion. Cl diffusion profiles comparing wet chemistry and microanalysis were generated from the quantitative map images of all samples. Mathematical fitting of normalized diffusion profiles generated with data from both instruments was used to determine Cl diffusivity (D sub a) in all samples subject to diffusion and was compared to the D sub a determined by wet chemistry. Analysis time for the largest areas studied took ~8 hr using the SEM and ~30 hr using the EPMA. Comparison of the Cl concentration data from both SEM and EPMA indicates that both methods consistently over-predict the Cl concentration at a given depth compared to wet chemistry. The over-prediction indicates that the standard sets used are not the most appropriate for these methods. This contention is well supported by the fact that using one of the diffusion-subject samples as a calibration standard yielded absolute concentrations for the other diffusion-subject similar to those determined by wet chemistry. Normalization and comparison of all data from wet chemistry and both instruments obviated the need for standards and generated D sub a's for all diffusion-subject samples within 6.8% and 4.4% error for Energy Dispersive Spectroscopy (EDS) and Wavelength Dispersive Spectroscopy (WDS) respectively compared to wet chemistry. These x-ray microanalysis methods therefore not only yield new data detailing the microscopic spatial distribution of Cl in cementitious materials, but could also be used in a streamlined approach to determine D sub a of Cl more rapidly than the wet chemical analysis techniques currently employed