Linewise kinetic Monte Carlo study of silicon dislocation dynamics

Abstract

We present a number of n-fold way kinetic Monte Carlo simulations of the glide motion of 90°partial dislocations in silicon. We undertake a survey of ratios of kink formation energy F k to kink migration barrier W m , over a range of temperatures and applied stresses. These simulations are compared with Hirth-Lothe theory and an extension to the Hirth-Lothe theory of Kawata and Ishiota. The latter is found to give the best description of the system. Using literature first principle values for the kink and soliton formation and migration energies, a model combining both strained bond and soliton mediated motion shows a negligible contribution to dislocation motion from the solitons. The high soliton pair creation barrier was limiting and a soliton mediated mechanism for dislocation motion would have to achieve thermal equilibrium concentration via impurity or point defect interaction to be effective. We also show that if this can be overcome solitons greatly increase the mobility of the dislocation, even without a binding energy between solitons and kinks. The simulation coded here is easily expandable to incorporate further dislocation line effects such as impurities at the line.

Dislocation motion in silicon: the shuffle-glide controversy revisited

“…Unlike the Duesbery and Joós model, such a glide / shuffle transition would then be explained by a change of mobility regimes, according to the average separation between thermal kinks. Such a scenario is supported by an interesting study by Scarle et al suggesting that the regime R 2 should be used for kink formation energies lower than 0.4-0.5 eV [29]. In this case, keeping 1.2 eV for the migration energy, the kink formation energy of therange 0.008-0.016µ.…”

Dislocation motion controlled by interactions with crystal lattice: modelling and experiments

“…The lower dislocation velocity at low stress in medium purity Si is accordingly the result of interactions with impurities, not the result of the transition between correlated and uncorrelated kink motions predicted by Moller. 11 The strict proportionality between velocity and stress is in agreement with the Hirth and Lothe theory 1 and with Scarle et al 12 but contrary to the estimation of Indenbaum et al 5 that predicts m equal to 3/4 or 7/8. Figure 5b shows results from different origins where impurity effects are considered to be negligible, because (i) measurements are from high purity crystals in the low stress range, and (ii) stresses are high enough to overcome pinning points in medium purity crystals at high stresses (which is attested by the rectilinear shape of moving dislocations in Ref.…”

Unraveling the temperature dependence of the yield strength in single-crystal tungsten using atomistically-informed crystal plasticity calculations

“…In the last two decades, computer simulation has unquestionably emerged as discipline capable of shedding light on these processes on a similar footing with experiments, providing physically-substantiated explanations across a range of temporal and spatial scales. These include the use and application of density-functional theory methods (Ventelon & Willaime, 2007;Ventelon et al, 2013;Weinberger et al, 2013;Dezerald et al, 2014Dezerald et al, , 2015, semi empirical atomistic calculations and molecular dynamics calculations Queyreau et al, 2011;Chang et al, 2001;Komanduri et al, 2001), kinetic Monte Carlo (Lin & Chrzan, 1999;Cai et al, 2002;Deo & Srolovitz, 2002;Scarle et al, 2004;Stukowski et al, 2015), and crystal plasticity (CP) (Qin & Bassani, 1992;Dao & Asaro, 1993;Brünig, 1997), to name but a few. In general, while there is no doubt that the intricacies associated with 1 2 x111y screw dislocation glide -including its thermally activated nature and deviations from Schmid law-cannot but be resolved using methods capable of atomistic resolution, one must recognize that, at the same time, flow is a phenomenon potentially involving statistically-significant amounts of dislocations and -as such-cannot be captured resorting to atomistic calculations only.…”