Ionic diffusion coefficients are estimated rapidly using electromigration tests. In this paper, electromigration tests are accurately simulated by numerically solving the Nemst-Planck (NP) equation (coupled with the electroneutrality condition (EN)) using the finite element method. Numerical simulations are validated against experimental data obtained elsewhere [E. Samson, J. Marchand, K.A. Snyder, Calculation of ionic diffusion coefficients on the basis of migration test results, Materials and Structures/Materiaux et Constructions 36 (257) (2003) 156-165., H. Friedmann, O. Amiri, A. Ait-Mokhtar, A direct method for determining chloride diffusion coefficient by using migration test, Cement and Concrete Research 34 (11) (2004) 1967-1973.]. It is shown that migration due to the non-linear electric potential completely overwhelms diffusion due to concentration gradients. The effects of different applied voltage differences and chloride source concentrations on estimations of chloride diffusion coefficients are explored. We show that the pore fluid within concrete and mortar specimens generally differs from the curing solution, lowering the apparent diffusion coefficient, primarily due to interactions of chloride ions with other ions in the pore fluid. We show that the variation of source chloride concentration strongly affects the estimation of diffusion coefficients in non-steady-state tests; however this effect vanishes under steady-state conditions. Most importantly, a comparison of diffusion coefficients obtained from sophisticated analyses (i.e., NP-EN) and a variety of commonly used simplifying methods to estimate chloride diffusion coefficients allows us to identify those methods and experimental conditions where both approaches deliver good estimates for chloride diffusion coefficients. Finally, we demonstrate why simultaneous use and monitoring of current density and fluxes are recommended for both the non-steady and steady-state migration tests. (c) 2007 Elsevier Ltd. All rights reserved.