Stretching the limits

My interest in plastic membrane mirrors began in the 1970s. They appeared to offer a solution for inexpensively visualizing infrared (IR) scenes and providing optics for a range of 3-D imaging systems. At that time the only plastic membrane mirrors available were adjustable shaving mirrors, in which a disc formed of finished plastic approximately 300 µm thick was circumferentially gripped between a pair of circular flat rings. A small rubber plunger evacuated air from behind the membrane, allowing air pressure difference to push the membrane back slightly to create a very shallow concave mirror that magnified the viewer's face.

The mirrors were only a few inches in diameter, with f/#'s ranging from 4 to 6. Unfortunately, I sought large-diameter, optically accurate stretchable membrane mirrors with deep curvatures and f/#'s of around 0.5. In the 1970s, the solution seemed a long way off.

Throughout the summer of 1979, I was working at the research labs of Ford Motor Co. (Dearborn, MI) to devise a method for imaging objects rotating at high speeds, which would allow engineers to monitor the heating patterns of rotating car brake discs, for example. Gordon Brown, Ray Wales, and I hatched an idea to inexpensively visualize IR emissions from warm stationary or rotating components. The catch was that on the stationary components, we wanted to visualize vibration patterns on cold steel car bodies, which required measuring temperature differences of 0.001°C.

Earlier in 1979 I had published a paper on visualizing the radiated heat from warm, rotating components. A sheet of liquid crystal (LC) was spun synchronously with the warm rotating component. A 12-in.-diameter, f/1.5 concave mirror focused the radiated heat patterns onto the LC material. When we illuminated the LC with white light, we revealed the heat patterns in real time, with the warmer bands appearing on the outer part of the image and the cooler bands near the center.

There was one hitch with applying this design to the Ford labs problem—the glass Porro prism it incorporated would not transmit the required IR wavelengths. Later, we devised a de-rotator that replaced the prism with membrane mirrors, making the device compatible with the IR spectral region, which allowed us to de-rotate the thermal image of the rotating component. The concave mirror then focused the image onto a sheet of stationary LC, providing excellent images. Theoretically, such a design could be built using conventional optics, but they would be expensive and would require a fixed curvature. The variable-curvature membrane mirrors would allow it to be used at different focal lengths in different applications.

Returning to Scotland in the fall of 1979, I began the search in earnest for a large-diameter, small f/# concave mirror that would enable us to build systems such as the all-mirror de-rotator. In IR studies, the heat produced in the mirror-focused zone is proportional to the reciprocal of the f/#². By using an f/0.5 mirror instead of an f/1.5 mirror, the amount of heat produced in the image is increased by (1.5/0.5)², or a factor of nine. The much larger diameter also collects more heat radiated from the component. By 1982 I had created 2-ft-diameter, f/0.7 stretchable plastic membrane mirrors. Membrane thickness varied between 10 and 150 µm. The material used was a bi-axially stretched nylon, also known as Mylar. By 1984 I was presenting sharp white light and IR visualized images formed with such mirrors-- images good enough to resolve fine detail, such as the individual hairs on the back of a man's hand.

In 1985 Scottish telescope-maker John Braithwaite designed a Newtonian telescope based on a 6-in. diameter membrane mirror. The primary mirror of the telescope could also be operated as an adjustable off-axis mirror, which eliminated the need for a secondary flat mirror that otherwise degraded telescope performance by blocking a large part of the target radiation reaching the image plane. The mirror also operated with a simple retro-reflector that passed the target image back through the optical system and removed all the aberrations of the optics, giving near-diffraction-limited performance.

Following the 1985 publications, mirrors were delivered to a Los Angeles division of TRW (Cleveland, OH), which developed a mirror that worked to diffraction-limited efficiency with pulsed laser light. In the summer of 1986, Arthur Guenther, then chief scientist at Kirtland Air Force Base (Albuquerque, NM), saw the mirrors in action at Glasgow. The Air Force was interested in the potential use of membrane optics for large-diameter telescope applications and is now doing its own research in this area.

By the early 1990s, Jim Fischbach of American Propylaea Corp. (Detroit, MI) investigated the membrane mirror with the goal of creating a large holographic optical element to be used to project stereo pairs directly into the eyes of an observer. The idea had its origins with John Logie Baird of Scotland. In 1943, researchers used a large lens to project stereo pairs into a viewer's eyes, with superb results. The problem is that very large lenses of small f/# are not available. The solution is to use an equivalent concave mirror, hence Fischbach's interest in the membrane mirrors.

In 1993 I met Tom Scott, then director of the Advance Design Studio at Ford Motor Co. (Dearborn, MI) and Bob Andrews, a design technologist, who applied the mirrors to 3-D imaging. Ford was interested in the possibility of creating 3-D displays that could be used for vehicle design, as well as for advertising purposes. Producing a volumetric image the size of a vehicle is impractical with standard optics; such a system would require very large-diameter lenses, which are expensive and difficult to make. A stretchable-membrane mirror, on the other hand, is lightweight, economical, and offers a variable focal length and a specular reflecting surface. Such mirrors do not suffer from chromatic aberration and have much higher resolution than Fresnel lenses. Between 1994 and 1998, the University of Strathclyde, funded by Ford, made optics 1.2-m in diameter and used them in novel 3-D displays with the variable focal length mirrors. Using a 1.2-m-diameter mirror with a 1550-mm focal length, for example, the university produced sharp, bright, full-size images viewable from an angle up to 22°.^1,2

More recently, Ethereal Technologies (Ann Arbor, MI) is commercializing the mirrors as a core component of special 3-D imaging systems. Strathclyde licensed the technology to the company, which now sells an autostereoscopic workstation based on the mirrors. In 2000, using video compression techniques developed by Gordon Mair of the University of Strathclyde's Transparent Telepresence Group, we sent live autostereoscopic images from Glasgow over the Internet to Ann Arbor, 4000 mi. away. My arm appeared to emerge from the system in Ann Arbor to shake hands with Professor Emmett Leith, a leading researcher in modern holography. The demonstration opens the Internet for glasses-free 3-D imaging, for companies ranging from advertising to live teleconferencing.

I completed this article while working on the Visible Human Project at the University of Michigan (Ann Arbor, MI), which provides detailed images of the human body to medical students via the Internet (see oemagazine, July 2001, page 26). The team, led by Brian Athey, is working with Ethereal's system to develop a real-time, immersive display system for the project. This is just another example of how plastic membrane mirrors will provide new imaging systems in the not-too-distant future. oe

Acknowledgments

I wish to thank friends and colleagues, Stuart McKay, Steven Mason, Les Mair, Douglas Brown, Gordon Mair, and Ian Montgomery for their support over the years, as well as Bob Andrews at Ethereal Technologies for his vision and deep commitment to commercializing the membrane mirrors.

References

1. S. McKay et al., Stereoscopic Displays and Virtual Reality Systems VI, Proc. SPIE 3639, pp. 122-131 (1999).

2. S. McKay et al., Projection Displays V, Proc. SPIE 3634, pp. 144-155 (1999).