Place: Euromat 2015, Warsaw, Poland
Carbon is an important alloying element in steel. Due to its low solubility in ferrite, it tends to segregate towards dislocations, forming cloud like pattern known as Cottrell atmospheres. The hardening observed during strain aging experiments is attributed to this effect, since the segregated impurities pin dislocations and reduce their mobility. Modeling this process on an atomistic length scale is difficult, since diffusion and dislocation mobility take place on different time scales. A novel method is presented by coupling Molecular Dynamics (MD) and Monte Carlo (MC) in a unified framework to bridge these time scales. First principle studies show that octahedral sites are energetically favorable location for carbon in ferrite. The link between MD-MC is established by initially placing virtual particles at octahedral sites of a bulk crystal. In the scope of MD, virtual particles mimic the behavior of carbon atoms by finding energetically favorable sites during plastic deformation, thus they are not restricted to octahedral sites. These particles do not exert a force on iron atoms, but get influenced by the force of surrounding iron atoms using a potential function similar to the Fe-C interaction. Diffusion is then modeled using the Metropolis MC algorithm in which the virtual particles act as sampling sites for the insertion of carbon atoms, thus the segregation of carbon can be modeled by sufficient large number of MC steps. The influence of carbon segregation on the dislocation mobility is studied using this framework. This coupling approach is extended to large scale system by an efficient parallelization of the Monte Carlo scheme.