Abstract.The use of high-Z plasma facing components requires intensive research in all areas, i.e. in plasma wall-interaction, in the physics of the confined plasma, diagnostic, and in material development. Only a few present day divertor tokamaks -mainly Alcator C-Mod and ASDEX Upgrade -gained experience with the refractory metals molybdenum and tungsten, respectively. ASDEX Upgrade was stepwise converted from graphite to tungsten PFCs and in parallel a reduction of the deuterium retention by almost a factor of ten has been observed due to the strong suppression of D co-deposition with carbon. The deuterium retained in W is in line with laboratory results. In order to diagnose W sources and the W content in the main plasma adequate spectroscopic methods had to be developed. As expected from the sputtering threshold of Mo and W, negligible erosion by the thermal divertor background plasma is found in ASDEX Upgrade and Alcator C-Mod under low temperature divertor conditions. However, erosion by fast particles and intrinsic impurities, which additionally might be accelerated in rectified electrical fields observed during ion cyclotron frequency heating, plays an important role. The Mo and W concentrations in the plasma centre are strongly affected by plasma transport and variations up to a factor of 50 are observed for similar influxes. However, it could be demonstrated that sawteeth and turbulent transport driven by central heating can suppress central accumulation. The inward transport of high-Z ions at the edge can be efficiently reduced by 'flushing' the pedestal region caused by frequent edge instabilities. Since with metal walls the edge radiation by low-Z impurities is reduced, it has to be substituted in a pure high-Z device by artificially injected low-Z impurities in order to keep the power load at an acceptable level. Experiments at ASDEX Upgrade suggest that a regime with benign erosion and favourable confinement can be achieved. Extrapolations to ITER and DEMO are difficult since the physics of plasma transport is not yet completely understood, the particle and energy fluxes are orders of magnitude higher and the technical boundary conditions in DEMO strongly differ from those of present day devices.