Search

Properties of basal and first order prism plane dislocations with Burgers vector 1/3 〈 1 1 2 ¯ 0 〉 in α-Zr have been investigated by computer simulation. Results for a recent interatomic potential (MA07) are assessed and compared with an older one (AWB95). The elastic constants have been calculated with the inner relaxations allowed for and the energy and vector of metastable stacking faults have been determined and compared with published ab initio estimates. The core of the screw dislocation spreads principally in the prism plane in the MA07 model, in contrast to basal plane dissociation in the AWB95 model, and the prism-to-basal ratio of the Peierls stress for the screw is 0.28 with the MA07 model, compared with 3.31 with the AWB95 model. Simulation of the dynamics of dislocation motion under applied stress reveal how the drag coefficient varies with slip system and temperature. The results for the MA07 model are consistent with the known slip geometry of Zr, suggesting that it offers significant advantages for large-scale atomic simulation of dislocation behaviour.

Dislocation segments with Burgers vector b = 〈1 0 0〉 are formed during deformation of body-centred-cubic (bcc) metals by the interaction between dislocations with b = 1/2〈1 1 1〉. Such segments are also created by reactions between dislocations and dislocation loops in irradiated bcc metals. The obstacle resistance produced by these segments on gliding dislocations is controlled by their mobility, which is determined in turn by the atomic structure of their cores. The core structure of a straight 〈1 0 0〉 edge dislocation is investigated here by atomic-scale computer simulation for α-iron using three different interatomic potentials. At low temperature the dislocation has a non-planar core consisting of two 1/2〈1 1 1〉 fractional dislocations with atomic disregistry spread on planes inclined to the main glide plane. Increasing temperature modifies this core structure and so reduces the critical applied shear stress for glide of the 〈1 0 0〉 dislocation. It is concluded that the response of the 〈1 0 0〉 edge dislocation to temperature or applied stress determines specific reaction pathways occurring between a moving dislocation and 1/2〈1 1 1〉 dislocation loops. The implications of this for plastic flow in unirradiated and irradiated ferritic materials are discussed and demonstrated by examples.

Atomic-level simulations are used to investigate the interaction of an edge dislocation with 〈1 0 0〉 interstitial dislocation loops in α-iron at 300 K. Dislocation reactions are studied systematically for different loop positions and Burgers vector orientations, and results are compared for two different interatomic potentials. Reactions are wide-ranging and complex, but can be described in terms of conventional dislocation reactions in which Burgers vector is conserved. The fraction of interstitials left behind after dislocation breakaway varies from 25 to 100%. The nature of the reactions requiring high applied stress for breakaway is identified. The obstacle strengths of 〈1 0 0〉 loops, 1/2〈1 1 1〉 loops and voids containing the same number (169) of point defects are compared. 〈1 0 0〉 loops with Burgers vector parallel to the dislocation glide plane are slightly stronger than 〈1 0 0〉 and 1/2〈1 1 1〉 loops with inclined Burgers vector: voids are about 30% weaker than the stronger loops. However, small voids are stronger than small 1/2〈1 1 1〉 loops. The complexity of some reactions and the variety of obstacle strengths poses a challenge for the development of continuum models of dislocation behaviour in irradiated iron.

Static and dynamic properties of the interstitial carbon atom and vacancy–carbon atom complexes in α-iron are modelled by a molecular dynamics (MD) method using a pair interatomic potential for the iron–carbon interaction and two different many-body potentials for the iron matrix. The diffusion parameters of an interstitial solute in iron are obtained by MD simulation for the first time. The binding energy and migration energy of a vacancy–carbon complex are also obtained: the complex is immobile and has higher energy for dissociation than the carbon atom migration energy. The results are compared with recent ab initio calculations and experimental data from the literature. Experimental data on the recovery stages of electron-irradiated Fe–C are analysed using rate theory equations and found to be consistent with the ab initio calculations for diffusion of a vacancy and its interaction with carbon atoms.

Precipitation of copper-rich clusters is a contribution to in-service hardening of some reactor pressure vessel ferritic steels. At temperatures less than 300 °C the precipitates are observed to be about 2 nm in diameter and not to coarsen, at least in the dose range from ∼10−3 to 10−2 dpa. As a result the hardening is close to a maximum. This phenomenon is studied here by computer simulations based on the “mean-field” approach for describing microstructural evolution in a binary Fe–Cu alloy. It is shown that the experimental data obtained from electron irradiated material and reactor-neutron irradiated steels have a stage of precipitate evolution intermediate between growth and coarsening. During this stage the size distribution of precipitates broadens while the number density and the mean size remain constant, which explains the observations. The role interstitial atom clusters produced in displacement cascades may have on the kinetics of copper precipitate coarsening is discussed.

The small, coherent BCC precipitates of copper that form during fast-neutron irradiation of ferritic steels are an important component in the irradiation hardening that occurs during service. The conventional explanation of hardening due to copper precipitates is given in terms of the Russell–Brown (M. Acta Metall. 20 (1972) 969) modulus hardening model, in which precipitates are treated as soft spots in iron. In the present paper, the core structure and energy of a screw dislocation are computed as it approaches a row of precipitates. The results indicate that the hardening observed in experiments is due to the effect of the screw dislocation core on the BCC copper structure rather than elastic interaction. This is a new precipitate strengthening effect. The increase in the flow stress is estimated from the interaction energy between the dislocation and the precipitate row, and the estimated value for precipitates of a size and spacing found in irradiated reactor pressure vessel steels is encouragingly close to that found experimentally.

Strengthening due to small coherent BCC precipitates of copper is an important component of in-service irradiation hardening of ferritic pressure-vessel steels. The dislocation effects involved are studied here by atomic-scale computer simulation. Many-body interatomic potentials for the Fe–Cu alloy system are used to investigate stacking-fault-energy surfaces and to simulate the atomic structure of the screw dislocation in both pure α-iron and the metastable BCC phase of copper. In iron, the core has the well-known three-fold form of atomic disregistry. In BCC copper, however, the core becomes delocalised by transformation of the copper as the lattice parameter is reduced to mimic the strain experienced by precipitates. Simulation of the screw dislocation threading through the centre of a BCC copper precipitate in an α-iron matrix shows that the extent of core delocalisation depends on precipitate size. The dislocation energy changes indicate a significant dislocation pinning effect due to this dislocation-induced precipitate transformation process.

Atomic-scale computer simulation techniques have been used to investigate the interaction of crystal dislocations with two interfaces in hexagonal-close-packed (h.c.p.) metals, namely the {10 1 2} twin boundary and a 〈1 2 10〉/90° tilt boundary which is incommensurate in the direction perpendicular to the tilt axis. Crystal dislocations are always found to be absorbed in the tilt boundary with concomitant reconstruction of their cores. In the twin boundary, a broader range of interactions is observed, including defect transmission from matrix to twin and decomposition in the interface into discrete defects. The easy generation of mobile twinning dislocations facilitates the latter behaviour. The simulations demonstrate that the core structures of localized interfacial defects exhibit preferred riser configurations. For the twin, the favoured structure is the “basal-on-prism” configuration, whereas risers in the tilt boundary resemble {10 1 2} twin forms. By comparing interaction processes in two interfaces, this investigation elucidates the role of crystallographic considerations and interfacial structure. It also illustrates that the core structure of interfacial defects can be complex and contributes significantly to total defect energy.

Atomic scale computer simulation of a model of the h.c.p. metal {alpha}-titanium is used to investigate the mobility of interfacial defects in response to applied shear stress. Interfacial defects in (10{bar 1}2) twins and a 90{degree} incommensurate tilt boundary are investigated. Defects with Burgers vector, b, parallel to their host interface can move conservatively in principle, but were found to be mobile only if their step height, h, is small. In such cases, particularly when the defects exhibit wide cores, the atomic shuffles involved in transferring atoms from sites of one crystal to the other are simple. Conversely, defects which exhibit large h generally have narrow cores and require complex shuffles for motion. Applied stress tends to cause core reconstruction and emission of partial dislocations trailing stacking faults from these defects. Defects with b inclined to the interface can move conservatively in response to applied shear stress through a climb-compensated mechanism in some circumstances. This mechanism can lead to limited mobility of defects in both types of interface studied, and involves the generation of additional glissile interfacial defects due to the stress concentrating effect of the riser of the initial defects. Activation of this mechanism is only feasible when themore » elementary mechanism of motion involves a small number of atoms shuffling from one crystal to the other. Unlike the case for defects with b parallel to the interface, this number is not simply related to h and can be effectively small even when h is relatively large.« less