Modeling and validation of chemical vapor deposition for tungsten fiber reinforced tungsten

In nuclear fusion reactors there are extreme conditions for plasma facing components. Currently, pure tungsten (W) is used to withstand the enormous heat and particle fluxes, neutron irradiation and plasma erosion. However, a significant drawback is that W is inherent brittle and thus can fail without warning. Therefore, W fiber-reinforced composites (W$_{f}$/W) are currently being developed. These can be produced by coating W fabrics via chemical vapor deposition (CVD). The aim of this work is to provide a profound and quantitative understanding of this process so that the material properties can be further improved. Models for the W-CVD process have been developed utilizing the commercial software COMSOL Multiphysics, and validated against experimental results. As a highlight a new description of the reaction kinetics was proposed solving controversies in literature with respect to the reaction order of the precursor WF$_{6}$ [1]. An increased W$_{f}$/W strength can be reached by higher relative density and finer W grains. Experimental CVD parameter studies and infiltration simulations showed that both can be achieved by operating at low WF$_{6}$ gas flow rates, high H$_{2}$ gas flow rates, medium total pressures and low temperatures. As lower temperatures increase the needed deposition time exponentially, 723–773 K are recommended. In addition, it is important to avoid WF$_{6}$ depletion within fiber inter-spaces, as this will lead to rapidly decreasing deposition rates, remaining pores, and thus to a reduced relative density. The minimum necessary WF$_{6}$ gas flow rate can be calculated with the developed model. The theoretically optimized CVD process parameters were applied experimentally to produce new bulk 15-layer W$_{f}$/W. In addition to finer grains and a higher relative density, further successes were a larger fiber volume fraction, a reduced WF$_{6}$ demand and a more uniform macro-scaled coating thickness. On the microscale, the new parameters resulted in such an uniform deposition that the optimization of the W fiber positions within the fabric (CVD substrate) was simplified based on geometric equations. However, in practice, pores can still remain at certain locations due to fiber positional deviations. Concepts for reducing these deviations and also for improving a continuous W$_{f}$/W production are presented in the Outlook chapter