Devices containing planar arrays of micromachined, electrostatically adjustable Fabry-Perot interferometers are undergoing development. These devices could be designed, for example, as color high-definition television displays, as larger flat-panel displays for indoor and outdoor entertainment and advertising, as filter arrays for spectroscopy, and as modulator arrays for optical computing and switching. In comparison with state-of-the-art flat-panel display devices based on liquid crystals, plasmas, and electroluminescence, the present devices offer potential advantages of high speed, insensitivity to changes in temperature, low power consumption, wide viewing angle, scalability, light weight, and long life.

A related concept of using two-stage, micromachined, electrostatically adjustable Fabry-Perot interferometers as rapidly tunable color filters and shutters was presented in "Micromachined Opto/Electro/Mechanical Systems" (NPO-19467), NASA Tech Briefs, Vol. 21, No. 3 (March 1997), page 50, and "Micromachined Tunable Optical Interference Filters" (NPO-19456), NASA Tech Briefs, Vol. 21, No. 3 (March 1997), page 111. The devices being developed according to the present concept are based on the same physical principles but differ in significant details of design and modes of operation.

Each of Three Micromachined Interferometers in a pixel would either transmit light at a resonant wavelengthli when relaxed at gap di, or else would not transmit when its mirrors were pulled together by voltage applied to electrostatic-deflection electrodes.

In a three-color television display device according to the present concept, each pixel would contain three micromachined, electrostatically adjustable Fabry-Perot interferometers, each serving as a modulator for light of one of three wavelengths (see figure). Each micromachined interferometer would contain two parallel, flat, partially transparent mirrors - one on a springy silicon nitride membrane and the other on a stationary glass substrate. The mirrors and the gap between them would constitute an optical cavity with resonant transmission peaks at wavelengths equal to integer submultiples of twice the size of the gap; that is, the interferometer would transmit most of the light at these wavelengths and reflect most of the light at other wavelengths.

The nominal size of the gap in each micromachined interferometer would be selected so that its resonant wavelength in the visible part of the spectrum was that of the desired color. The display panel would be illuminated with white light from its back side (the lower side in the figure). Optionally, color filters could be formed on the back side registered with the corresponding interferometers to provide additional selectivity for greater purity of the colors.

When voltage was not applied to the electrostatic-deflection electrodes, the springy silicon nitride membrane in each interferometer would maintain the nominal gap and therefore light of the nominal wavelength would pass through to the front (top in the figure) side, where it would be seen. When voltage was applied to the electrostatic-deflection electrodes of a given color interferometer in a given pixel, the spring force would be overcome and the two mirrors drawn together; this would eliminate the resonant gap, causing the two mirrors to act as ordinary mirrors so that light would not pass through to the front. The net effect would be that each interferometer would act as a light valve or modulator for its assigned color. Thus, by opening each light valve for a specified fraction of each image-repetition cycle, one could mix specified proportions of each color. Since the viewer's eye could not spatially resolve the individual interferometers or temporally resolve the individual flashes of light, the viewer would get the impression of a desired composite color emanating from the pixel.

This work was done by Tony K. T. Tang, Linda M. Miller, Michael H. Hecht, and Judith A. Podosek of Caltech for NASA's Jet Propulsion Laboratory. NPO-19527