In ITER, a loss of vacuum accident is likely to re-suspend particles produced and deposited on the torus walls. The thesis purpose is to provide reliable experimental data to develop and validate a re-suspension model, taking into account the low pressure and predicting the amount of particles mobilized during such event. Experiments were carried out as a function of key parameters — particle diameter, friction velocity and density of the atmosphere — entering in the re-suspension mechanism. The size distribution of tungsten particles produced in tokamaks shows diameters from 0.1 to 10 μm. The experimental data of tungsten particle monolayer deposits re-suspension are in good agreement with the Rock’n Roll model when the particles are only mobilised by the fluid. However, a mobilisation of fine particles by clustering, no taken into account in re-suspension models, was highlighted. My experiments have also revealed the importance of the multilayers deposit structure (porosity and thickness) in the re-suspension mechanism with, here again, a mobilisation by particle clustering. The effect of low pressure in the re-suspension mechanism was studied between 10 and 1,000 mbar and shows a friction reduction in the airflow forces at low pressure. Finally, measurements of the MES phenomenon kinetics highlighted three distinct regions with a short-term effect, a long-term effect and a transition region which must be taken into account in a MES calculation.