Two material systems, both signi cant as structural materials in nuclear reactors are modelled using atomic-scale simulations. The first is beryllium metal, which in fusion and fission reactors, is used as a neutron reflector, but in fusion applications is also being considered as a neutron multiplier and a plasma facing material. The low neutron absorption cross-section and high neutron scattering cross-section of beryllium makes it suitable in these roles. The second system is ZrO2, which forms on the surface of zirconium-based fuel cladding during operation in fi ssion reactors. ZrO2 acts as a barrier layer between the cladding metal and the coolant water. Dopants in the cladding metal inevitably end up in the ZrO2 as a consequence of its formation
due to oxidation of the cladding metal and can impact its performance as the barrier layer and
in turn the expected cladding lifetime. Zirconium is also a low absorption cross-section element.
For the beryllium system, molecular dynamics (MD) damage cascade simulations were performed.
Displacement damage in beryllium was predicted as a function of temperature and energy. For each temperature and energy pair, 10 simulations were performed with a predetermined set of PKA directions. Branch-like damage cascades were observed, with defects being formed primarily on the cascade backbone. The number of residual defects remaining after 1 ps increased linearly with primary knock-on atom (PKA) energy from 0.5 keV to 2.5 keV, while the extent of residual damage was largely temperature independent from 300 K to 1100 K.
This was attributed to the mechanism by which displaced atoms within the disordered damage
region recombine and are thus are unlikely to survive as residual damage. The same simulation
model was used to predict the directionally averaged probability of displacement as a function
of displacement energy, P(E_PKA), and thereby the threshold displacement energy at which the
probability for displacement is 100%, E_d(1.0) = 105 eV. There is an excellent correspondence between
the Kinchin-Pease (K-P) prediction using E_d = E_d(1.0) and the number of residual defects
remaining after the initial recovery phase. Also, by utilising P(E_PKA), a modi cation to the
K-P model that utilises the probabilistic nature of threshold displacement is proposed.
The behaviour of dopant ions was investigated in monoclinic and tetragonal ZrO2. Formation
energies of dopant-containing as well as intrinsic defects were calculated via density functional
theory (DFT). The formation energies were then used to construct defect concentration diagrams
(i.e. Brouwer diagrams), which are a function of oxygen partial pressure PO2 . All the dopant elements considered (Fe, Ni, Cr, Cu and Sb), were found to bind strongly to oxygen vacancies in tetragonal ZrO2. At low PO2 , expected in the vicinity of the oxide metal interface, a cluster consisting of an oxygen vacancy adjacent to a charge neutral Ni0 atom was identifi ed as the most populous cluster. Further simulations show that a hydrogen molecule will dissociate in the vicinity of this cluster. No other cluster containing any of the ve dopants is both suffciently populous and acts in this manner. This is proposed to be the mechanism by which Ni is detrimental to hydrogen pickup in Zr alloys. Consequently, neither Cu nor Sb are expected to increase hydrogen pickup in the same way. Sb was identi ed as a particularly suitable alloy
addition compared to Cu as potential alloying element in Zr-based cladding because Cu in ZrO2
leads to an increase in oxygen vacancy concentration whereas Sb does not.