A training effect on electrical properties in nanoscale BiFeO3

We report our observation of the training effect on dc electrical properties in a nanochain of BiFeO3 as a result of large scale migration of defects under the combined influence of electric field and Joule heating. We show that an optimum number of cycles of electric field within the range zero to similar to 1.0 MV cm(-1) across a temperature range 80-300 K helps in reaching the stable state via a glass-transition-like process in the defect structure. Further treatment does not give rise to any substantial modification. We conclude that such a training effect is ubiquitous in pristine nanowires or chains of oxides and needs to be addressed for applications in nanoelectronic devices.