Blazed binary optics, from pc to plastic
Richard D. Rallison
Scott R. Schicker
Stephen E. Bialkowski, Ph.D.
Dept of Chem and Biochem
Utah State University
Logan, Utah 84322
A zone plate with 250 zones has been computed on a 386 machine and printed on 100 sheets of plain paper in a NEC LC890 printer. It was computed in Postscript language with 8 levels of grey in each zone and measured 75 inches in diameter before photoreduction to millimeter dimensions in a Nikon FA camera using Kodak 5052 TMX film. The reductions were made from 50 to 300 times to test the range of system resolution, fine grain Illford holographic film was substituted to test the camera MTF. Three micron layers of resist were used to get maximum phase shift and embossing masters of epoxy and nickel were made and evaluated. The final elements are embossed or cast in a variety of plastics and epoxies on glass, metal or semiconductor substrates. The final copies are efficient micro lenses and gratings useful for fabricating beam deflectors and a variety of stable optical interconnects. Resolutions are not as high as electron beam or laser scanning techniques but equipment and fabrication costs are low.
We wrote a program named ZONE2 that will compute and print round or elliptical geometric zone plates with up to 20 shades of grey between zones. A second program named LINES writes linear gratings of the same sort, both generate postscript files that can be scaled as large as the computer system will handle. In our case the practicle upper limit was a 10 megabyte file that printed 250 zones times 8 shades of grey spread over 100 pages. The number of zones per page was found to be limited by the printer artifacts related to printer resolution and the computing of shades of grey in the postscript language. The number of grey shades is similarly limited to about 11 distinguishable shades in the inner zones and perhaps 6 to 8 shades in the outer zones.
Geometric zone plates are free of aberrations only for f #s larger than 10 so the restriction to 250 zones means we are limited to a lens that is 10 mm in diameter with a focal length of 100 mm at a wavelength of 500 nm. If we had added a term and made the calculations for an interferometric zone plate we could have gone to an f# of 4 without aberrations using 250 zones in a diameter of 4 mm and a focal length of 16 mm. For some applications we could ignore the spherical aberrations of the geometric zone plate used at lower f #s and we made some in the f # 4 range for examination. Optics in this range require resolutions better than 250 l/mm at their outer limits. The camera was fitted with an f/2.8 Micro-NIKKOR 50mm lens which could reasonably resolve 4 micron details over a small angular field.
The camera had to be loaded with individual frames of ILLFORD holographic film to insure a large enough film MTF, then each piece was hand processed in DEKTOL developer which yielded good grey scale. Shooting was done outside on a sunny day at distances of between 12 and 100 feet. The zone plate was cut and pieced together on an 8 by 7 foot substrate made from 2 inch thick urethane foam laminated to smooth white paneling.
For maximum thickness we chose SHIPLEY #1400-37 PhotoResist spun onto 2.5 X 2.5 inch float glass squares for making the surface-relief copies of the Silver Halide masters. We had no previous experience with this material and this is what we found would work.
1) We set up a Gyrex IR oven with the low pre-heat on, the main heater control set to position seven, and the conveyer speed set to slowest. We washed the substrates, rinsed them in D.I. water, and dried them in a class 100 clean hood. When the substrates appeared dry,we ran them through the I.R. oven two or three times to dehydrate them and stored them in a clean hood until ready to coat.
2) The oven is then cooled down and set up for the "Soft Bake" step. The pre-heater is turned off, the heater is set to one,(130 to 140 degrees F) and the conveyer is set to its slowest speed ( 6 minutes). A soft bake for the photoresist is accomplished by running it through the oven only twice in succession. Over-baked resist does not react well to light.
3) We use a multiple discreet speed DC spin-coater with a small suction cup for applying the photoresist. The speed control is set to get approximately 900 revs per minute as measured by a mechanical tachometer. The spin coater box is lined with aluminum foil and set up on a class 100 clean bench. A substrate is centered on a suction cup and 3ml of the resist is applied to the substrate. The assembly is moved so as to spread the resist evenly over the surface to be coated. The suction cup is inserted into the spin-coater and spun for three minutes.This produced a dry 6 micron thick coating as measured by a .001 inch resolution dial indicator, ( 2.5 microns per division).
4) After a substrate is coated, immediately run it through the oven for the soft bake step. It is advisable to establish a routine so that residual solvents are uniformly distributed and sensitivity is then consistent and uniform across the substrate and from plate to plate.
5) To make a contact copy of a silver-halide master, we place the master with the emulsion side up onto the photoresist and weight the master down with a half inch thick slab of suprasil. Next we place a non-reflective mask over the suprasil to prevent edge scattering and reflections. Lastly, we expose the whole works to a standard 175 WATT mercury vapor yard light from which the outer glass globe has been removed to allow the full spectrum of UV to escape. Thirty minutes at a distance of one foot worked quite well for us and the lamp had to be fan cooled at that distance or it would melt the plastic.
6) The development process consists of washing the exposed photoresist off of the substrate with a dilute solution of MicroPosit 351 developer ( more commonly known as sodium hydroxide ). We used a solution of three parts water with one part of the MicroPosit 351 Developer Concentrate. Development is carried out in a small tank by uniformly immersing the plate and gently agitating only two times back and forth and then letting the plate lean against the side of the tank film side down. After developing the exposed resist for four minutes, we then rinsed it in a dilute solution of Kodak stop-bath ( 3 ml per 500 ml total volume ) and then De Ionized water for ten minutes. We tried using photo-flow in the rinse water but it left a residue. Dry the finished gratings in a clean hood under a gentle flow of air. DO NOT blow it dry with high pressure air, the grating is easily blown apart at this stage.
7) When we tried hard-baking the gratings in the Gyrex oven we succeeded in melting the resist. We now skip the hard-bake step as it does not seem to be necessary for electroplating or replicating with epoxy.
We exported some resist masters to Dazzle Enterprizes for electro plating but we are preparing to do our own plating in the future. Electro plating is necessary to replicate with full depth onto substrates using UV epoxies or thin plastic films. We received 8 shims from two different resist masters made at 28 microns from Dazzle. The shims were made in 4 different thicknesses, 2, 4, 6, and 8 mils thick and were not as flat as we had hoped. Casting against them was carried out with UV epoxies and a nip roller with good results when using the 4 mil shims.
Copies were also easily embossed into various plastics using a solvent and a Foster replicator. The Foster replicator is a hand cranked ringer, commonly used to ring out wet articles of clothing.
Last modified on 6/3/99