Photorefractive 3-D disks for optical data storage and artificial neural networks

Citation

Li, Hsin-Yu Sidney (1994) Photorefractive 3-D disks for optical data storage and artificial neural networks. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/wx1q-v852. https://resolver.caltech.edu/CaltechETD:etd-03012004-161831

Abstract

This thesis is on the application of 3-D photorefractive crystals disks for holographic optical data storage and optical neural networks. Chapter 1 gives some introductory background and motivation for the materials given in this thesis. In Chapter 2, the coupled-mode analysis and Born's approximation in anisotropic crystals is reviewed. The results are similar to that of isotropic materials. However, there are approximations that are often neglected in the literature. Chapter 3 starts with the description of the holographic 3-D disk for data storage, and analyzes the various alignment errors and tolerance problems for a 3D disk system. Of particular interest is the effects in image reconstruction caused by rotational angle error. An optimum configuration is found that minimizes this error. Chapter 4 examines the data storage density of 3-D disks and volume holographic storage systems that utilize wavelength/angle and spatial multiplexing. The maximum storage density and the geometry that achieves this density is derived. Chapter 5 discusses the diffraction efficiency of 3-D disks fabricated with photorefractive crystals. Practical geometries and crystal orientations for achieving maximum uniform diffraction efficiency are given and compared to the maximum obtainable diffraction efficiencies using arbitrary cut crystals. Experimental results are shown. Also derived in this chapter are the double grating effect from crystal anisotropy, and the optimum configuration for getting maximum diffraction efficiency using the 90 degree recording geometry. The Kuhktarev band-transport model of the photorefractive effect is examined briefly with emphasis on the anisotropy of the material. The proper expression for the permittivity term in the space-charge field formula is derived. Chapter 6 gives an example of an optical neural network that uses photorefractive crystals. It is the real time face-recognition system. The setup and experiments are described. Some properties of volume holographic correlators are given in the Appendix.

Thesis Files