Molecular dynamics simulations of interactions between hydrogen and fusion-relevant materials

Research output: ThesisDoctoral thesis 1 (Research UU / Graduation UU)


In a thermonuclear reactor fusion between hydrogen isotopes takes place, producing helium and energy. The so-called divertor is the part of the fusion reactor vessel where the plasma is neutralized in order to exhaust the helium. The surface plates of the divertor are subjected to high heat loads and high fluxes of energetic hydrogen and helium. In the next generation fusion device - the tokamak ITER - the expected conditions at the plates are particle fluxes exceeding 1e24 per second and square metre, particle energies ranging from 1 to 100 eV and an average heat load of 10 MW per square metre. Two materials have been identified as candidates for the ITER divertor plates: carbon and tungsten. Since there are currently no fusion devices that can create these harsh conditions, it is unknown how the materials will behave in terms of erosion and hydrogen retention. To gain more insight in the physical processes under these conditions molecular dynamics simulations have been conducted. Since diamond has been proposed as possible plasma facing material, we have studied erosion and hydrogen retention in diamond and amorphous hydrogenated carbon (a-C:H). As in experiments, diamond shows a lower erosion yield than a-C:H, however the hydrogen retention in diamond is much larger than in a-C:H and also hardly depending on the substrate temperature. This implies that simple heating of the surface is not sufficient to retrieve the hydrogen from diamond material, whereas a-C:H readily releases the retained hydrogen. So, in spite of the higher erosion yield carbon material other than diamond seems more suitable. Experiments suggest that the erosion yield of carbon material decreases with increasing flux. This was studied in our simulations. The results show no flux dependency, suggesting that the observed reduction is not a material property but is caused by external factors as, for example, redeposition of the erosion products. Our study of the redeposition showed that the sticking probability of small hydrocarbons is highest on material previously subjected to the highest hydrogen flux. This result suggests that redeposition is more effective under high than under low hydrogen fluxes, partly explaining the experimentally observed reduction in the carbon erosion yield. Lastly, we studied amorphous tungsten carbide. Amorphous material with three different carbon percentages (15, 50 and 95%) was subjected to deuterium bombardment and the resulting erosion and deuterium retention was analysed. The 95% carbon sample behaves like doped carbon, the carbon erosion yield is reduced and no tungsten is eroded. Segregation of the materials was observed, resulting in an accumulation of tungsten at the surface. The hydrogen retention was similar to a-C:H. The 15% carbon sample showed no significant erosion or retention. The most interesting was the 50% sample. Here deuterium bubbles formed that burst open after sufficiently long bombardment, thereby removing both carbon and tungsten from the surface. In the context of ITER our MD simulations suggest that tungsten is the better suited material since both the erosion and the hydrogen retention are significantly lower than for carbon.
Original languageUndefined/Unknown
QualificationDoctor of Philosophy
Awarding Institution
  • Utrecht University
  • Goedheer, W.J., Primary supervisor
  • Kleyn, A.W., Supervisor, External person
Award date24 Feb 2010
Print ISBNs978-90-393-5282-3
Publication statusPublished - 24 Feb 2010

Cite this