Motion of 180° Domain Walls in BaTiO3 under the Application of a Train of Voltage Pulses

Publisher Website
Google Scholar
Abstract
A study has been made of the sidewise motion of 180° domain walls in single‐crystal BaTiO3 under the application of a train of voltage pulses. Measurements of the sidewise wall velocity as a function of the pulse amplitude, pulse width, and pulse repetition rate can be explained in terms of a surface layer model recently proposed for BaTiO3. These data constitute the first experimental verification of some of the important phenomena consequent on the model. This model assumes that a surface charge density of 2Ps, where Ps is the spontaneous polarization, is deposited immediately behind an advancing 180° domain boundary on the interface between the bulk ferroelectric material and the anomalous surface layer. This charge distribution gives rise to an electric field Eb′ in the bulk of the material with a component opposite in direction to the field applied to move the wall. It is found that relaxation of Eb′ begins in about 0.3 msec and continues for times in excess of 1 sec. This relaxation is not due to a simple, single time‐constant, mechanism. Any fast relaxation, if present, must take place in times less than 1 μsec. After a 180° wall is at rest sufficiently long so that relaxation of Eb′ has occurred, application of an external applied field results in a wall velocity which depends on the magnitude of the wall displacement. In particular, the wall velocity decreases (due to Eb′ which increases) with increasing wall displacement. Analyses of the data show that the lateral or sidewise range of the field from the elements of charge which give rise to Eb′ is of the order of 2000 A so that once the wall has moved this distance, the effect of Eb′ at the wall, and hence the wall velocity, becomes constant. With thin crystals in the absence of an external applied field, Eb′ gives rise to wall displacements of the order of 100 A in a direction opposite to that of the wall motion which produced Eb′. The decrease in wall velocity observed when the wall displacement per pulse is of the order of a unit cell is consistent with a nucleation‐controlled mechanism proposed earlier for the wall motion. Pulse data on crystals with different impurity concentrations and electrode materials show that the relaxation of Eb′ and the lateral range r, and hence the surface layers inferred from these data, are not markedly dependent on these parameters. Other data on the dc wall velocity indicate that for a velocity of 10−3 cm/sec, the potential drop across the surface layer is 1.0±0.1 v, and that it is independent of the crystal dopings and the electrode materials tested. These data indicate that the average field in the bulk required for a given wall velocity is not a constant independent of the type of electrode or the impurity concentration. The approximate constancy of the measured characteristics of the surface layers makes the dependence of the wall velocity on the impurity content and the electrode material difficult to explain in terms of the model considered.