International Journal of Modern Physics B · 1996 · 61 citations · 0 references
Natural SciencesMaterials SciencePhysicsComputational Materials ScienceCondensed Matter PhysicsElectronic MaterialsCondensed Matter SimulationsDefect ModelingDefects In Electronic MaterialsMaterial ModelingPerovskite MaterialsPhase TransitionsLattice ModelsPotts ModelRandom Fields
A random field Potts model is used to establish the spatial relationship between the nanoscale distribution of charged chemical defects and nanoscale polar domains for the perovskite-based relaxor materials lead magnesium niobate (PMN) and lead scandium tantalate (PST). The random fields are not set stochastically but are determined initially by the distribution of B-site cations (Mg,Nb) or (Sc,Ta) generated by Monte Carlo NNNI-model simulations for the chemical defects. An appropriate random field Potts model is derived and algorithms developed for a 2D lattice. It is shown that the local fields are strongly correlated with the chemical domain walls and that polar nanodomains tend to nucleate about the latter. The evolution of the polar domains as a function of decreasing temperature is simulated for the two cases of PMN and PST. The dynamics of the polar clusters is also discussed. The results are used in an accompanying paper (H. Qian and L.A. Bursill, Int. J. Mod. Phys.B10, 2007, 1996) in order to predict the temperature and frequency dependence of the dielectric response of these two relaxors.