The hydrogen diffusion process in ferritic steel during thermal desorption tests was simulated using the finite element method based on the two-dimensional diffusion-trapping coupled model. This model was first verified by experimental data to obtain a physically meaningful combination of trap/lattice parameters. Then, the effect of specimen geometry was studied by varying the height of cylindrical specimens with other parameters fixed at constant values. Simulation of desorption spectra with different specimen geometries indicates that the measurement of hydrogen concentration is not affected by the change in specimen geometry due to the mass conservation law, for original thermal desorption spectra (TDS), which are, however, unlikely to be detected in traditional experiments due to the necessity of specimen transfer procedures. Considering the hydrogen escape during rest time (specimen preparation/transfer/evacuation), the measured TDS curves are expected to be strongly dependent on the specimen geometry. The effect of specimen geometry on desorption spectra is more pronounced for smaller specimens, resulting in the dramatic decrease in peak flux and the increased error of Kissinger method in the determination of trap deactivation energy. The present study may contribute to better understanding and more reliable interpretation of the TDS curves by considering the size effect.
Keywords: diffusion; hydrogen; simulation; thermal desorption spectroscopy; trapping.