• Energy Research

A novel identification technique of hydrogen transport parameters using FESTIM (Finite Element Simulation of Tritium In Materials) has been demonstrated. FESTIM is a finite element code developed with FEniCS performing hydrogen transport simulations. The trapping parameters (detrapping energies and trap densities) are identified for various materials (Tungsten, Aluminium, EUROFER and Beryllium) by automatically reproducing thermo-desorption experiments. Several optimisation algorithms are tested and the Nelder–Mead algorithm shows the best efficiency. An optimisation test problem with five free parameters took only a few hours to solve whereas optimisation cases with two free parameters took a few minutes. Limitations of this technique are shown and discussed.

Until recently the EIRENE suite handled hydrogen chemistry (i.e. H atoms, H2 molecules, and H2+ molecular ions and their isotopomers). Extensions also exist for hydrocarbon chemistry, including CxHy species (H is any hydrogen isotope), but no kinetic scheme including NHx chemistry had been implemented. After a benchmarking of available schemes and kinetic data, a reduced scheme including 50 reactive processes is presented here, implying a large set of N-bearing species (N, N2, N2+, NHx radicals and ions). This mechanism is a part of a more complete scheme of 130 processes also including the metastable states of N2 and N. In a first time, this scheme has been validated by comparison with independent experimental works. In a second time the scheme was used in Eirene suite to model a type-case. First results confirm that under tokamak conditions, ammonia is produced in significative amount after N2 injection in the sub-divertor volume. Keywords: Modeling, Edge-plasma, Ammonia formation, Tokamak

In order to simulate hydrogen charging and discharging cycles of mechanically loaded structures full 3D Macroscopic Rate Equation (MRE) modelling is proposed based on a finite element method (FEM). The model, implemented in the 3DS Abaqus software, uses a generalized transport equation, which accounts for mechanical fields, hydrogen transport and trapping, and their evolution with time. The influence of a-priori known thermal field has also been included. To ensure the solution convergence and the numerical stability, the trapping kinetic is introduced by using an approximation of the analytical solution the McNabb and Foster equation. Comparisons with a relevant 1D MRE code and with thermal programmed desorption (TPD) experimental results are performed on a 1D configuration to validate the model. Next, the model is used to simulate the tritium diffusion and trapping in a 2D geometry of interest in the upper plug of ITER tokamak, and results of tritium inventory are compared with an equivalent 1D calculation. Keywords: Hydrogen, Kinetic trapping, Modelling, Finite elements, Abaqus, User subroutine, Macroscopic Rate Equations

Diffusion and trapping of hydrogen isotopes in fusion materials need to be fully described in order to evaluate permeation and retention in fusion reactors walls and breeding blankets. Hydrogen gas permeation experiments have been conducted on Eurofer97 with pressures ranging from 101to 105Pa and temperatures between 473 K and 673 K, resulting in solubility K(T) (mol m−3 Pa−12)= 1.76⋅10−1exp(−0.27(eV)kBT), diffusivity D(T) (m2 s−1) =2.52⋅10−7exp(−0.16(eV)kBT)and permeability Φ(T) (mol m−1 Pa−12s−1) =4.43⋅10−8exp(−0.43(eV)kBT). Trapping parameters have been investigated using thermal desorption spectrometry of deuterium-loaded samples coupled with parametric optimization, leading to detrapping energies Edt,1=0.51eV, Edt,2=1.27eV, Edt,3=1.65eVand densities Nt,1=6.01⋅1025 m−3, Nt,2=6.44⋅1022 m−3, Nt,3=3.88⋅1023 m−3. This parametric optimization is performed using a kinetic surface model: the contribution of this model is compared to the results given by solubility and recombination rate models.

Experiments were performed in a low pressure-high density plasma reactor in order to study the impact of hydrogen retention in aluminium under plasma conditions. Microscopy scans of the surface were performed before and after 1h plasma exposure (fluence 6.1 ×1023ions/m2) where it is seen that blisters start to nucleate at the grain boundaries. Investigation on blister growth kinetics was performed for fluences ranging between 6 ×1023 and 3.7 ×1024ions/m2. The evolution of the characteristic size of the projected area was also analyzed. Finally, a macroscopic rate equations (MRE) code was used to simulate hydrogen retention and diffusion in Al and bubble growth in the bulk was simulated using experimental results. This model was also used to simulate these phenomena in Be and compare its behavior with respect to Al.

Aluminum samples have been exposed to a hydrogen plasma generated by a low-pressure – high-density microwave reactor. Aluminum has been chosen as a surrogate for Beryllium. The fluence was kept below 4 × 1024 ions/m2, in order to study the first steps of nucleation and growth of surface and bulk defects, i.e. blisters and bubbles. Experimental analyzes and macroscopic rate equation (MRE) modeling on poly- and single- crystals were made to investigate the role played by grains boundaries in the hydrogen retention. Temperature programmed desorption (TPD) on Al poly-crystals revealed the production of aluminum hydrides (alanes) as majority species in the desorption flux. Comparison of microscopy observations for three different single-crystal orientations (〈100〉, 〈110〉 and 〈111〉) allowed to determine preferential orientations able to attenuate the formation of blisters. Keywords: Blistering, Aluminum, Hydrogen inventory