Effects of applied voltage on hologram writing in lithium niobate

Abstract

The effects of applying voltages across the crystal during a repeated cycle of hologram writing followed by simultaneous reading and erasure have been studied with care to avoid errors due to multiple reflections. The results are of significance in connection with improving the efficiency in applications of lithium niobate as the hologram storage material in read‐write optical memory systems. They also provide information on the physics of the hologram storage process. It was found that large‐scale space‐charge fields, which are produced by the optically induced redistribution of electrons among traps, play an important role in determining the buildup of the sinusoidal space‐charge fields which produce the hologram. The large‐scale space‐charge fields cause the diffraction efficiency produced by a given exposure to depend on the voltage applied during the previous write‐read/erase cycles as well as on the voltage applied during the current cycle. Increased efficiency can be obtained by applying a different voltage during erase and write phases. Also, it is found that the efficiency of hologram writing depends on the fraction of the crystal illuminated. The development of these dc fields is modelled and compared with results from ellipsometric experiments. It is found that the effects of applied voltage are generally in agreement with hologram formation by the combined action of the new bulk photovoltaic effect of Glass et al., drift of photoreleased electrons in applied and space‐charge fields, and possibly some diffusion. The term κ₁αI (where κ₁ depends on the dopant, α is the absorption, and I is the light intensity), which Glass et al. used to describe the contribution to the electron current density provided by the bulk photovoltaic effect, is shown to be equivalent to what would be produced by a ’’virtual’’ built‐in field, provided the drift length was small. However, it does not entirely represent the behavior expected for their proposed model, and estimates of the virtual field are rather scattered. Intervalence transfer may occur. It is shown pyroelectric built‐in fields may be present under some circumstances but were not observed even under favorable conditions.