Abstract
Vector matrix multipliers have figured prominently in optical processing architectures because the minimal processing power requirements and the global communication needs of these devices dovetail nicely with the capabilities of optics. We describe an architecture for implementing vector matrix multiplication in integrated optics. The matrix is recorded in waveguiding lithium niobate by using unguided light to write photorefractive holograms. The matrix is read out by a vector of guided beams which are diffracted into an output vector. Crosstalk between guided beams is avoided due to the Bragg matching constraint for volume holograms. The diffraction of guided beams via gratings written with unguided light is accomplished using the rotational degeneracy of Bragg matched beams about the grating wave vector. By relaxing slightly the Bragg constraint it is possible to write an N × N matrix using a signal beam modulated by a mask of N2 pinholes and a single reference beam. We present preliminary experimental results concerning multiple holograms written in waveguides from above the waveguiding plane.
© 1987 Optical Society of America
PDF ArticleMore Like This
Gregory Gheen
THN1 OSA Annual Meeting (FIO) 1987
Q. W. Song, M. C. Lee, and P. J. Talbot
FBB6 OSA Annual Meeting (FIO) 1991
L. Scharf, W. T. Cathey, and K. M. Johnson
MD7 Optical Computing (IP) 1987