Nanoscale domain wall dynamics in ferroelectric thin films: effects of electro-mechanical field interactions

Ferroelectric oxide thin films are currently used in ultra high density non-volatile memories (FeRAM) and Nano/Micro-Electro-Mechanical Systems (NEMS/MEMS). In all these application, the functional property is determined by the ferroelectric and ferroelastic domain wall movement. Moreover, commercially developed systems are based on polycrystalline (or textured) ferroelectric thin films and therefore their performance rely heavily on the microstructural features such as orientation, grain size, grain boundary contribution etc. Furthermore, as the thin films are patterned onto the substrate for any device fabrication, the film-substrate interface defects such as lattice misfit, dislocations etc., highly influence the domain wall behavior.This dissertation, investigates the nanoscale domain switching behavior in polycrystalline perovskite lead zirconate titanate, Pb(ZrxTi(l-x)03 (PZT) ferroelectric thin films. Systematic Piezoresponse Force Microscopy (PFM) studies provide direct visual evidence of the complex interplay between electrical and mechanical fields in a polycrystalline system, which causes effects such as correlated switching between grains, ferroelastic domain switching, inhomogeneous piezostrain profiles and domain pinning on very minute length scales. Furthermore, the grain to grain long range interaction and ensuing collective dynamics in the domain switching behavior have been investigated using the time resolved PFM and Switching Spectroscopy PFM (SSPFM). Finite element method (FEM) has been employed to quantify the local ferroelectric interaction and assess the several possible switching mechanisms. The experiments find that of the three possible switching mechanisms, namely, direct electromechanical coupling, local built-in electric field and strain, and grain boundary electrostatic charges, the last one is the dominant mechanism.Having studied in detail the nanoscale domain wall behavior, we are now able to control and engineer the domain behavior