There are two classes of optical manufacturing methods: (1) the traditional grind and polish method that leads to an original optical element and (2) replication techniques that yield identical copies of a master. The various process steps in the two approaches are described to illustrate the fundamental differences in the methods. An understanding of these processes often is helpful for selecting the most appropriate manufacturing method.

No matter how a high precision optical element is manufactured, "replication" most likely offers a lower cost method to produce it in quantity. All parts made from the same master are essentially matched sets and identical in all their characteristics, a feat which would be very expensive to attempt in any other manufacturing method. Any substrate material which will satisfy the mechanical design requirements, can be optically surfaced by thin film replication. From a dozen space flight ultra light weight telescopes to high production military components, to a steady flow of commercial and instrument quality mirrors, "replication" has proven its cost effectiveness and overall viability. This paper will discuss a cross-section of reflective components which are representative of the thousands used every year by the optical industry.

In recent years, injection molding technology has become increasingly sophisticated, making possible the production of precision polymer optical systems. This technology continues to advance, and requires a corresnonding advancement in optomechanical design techniques, materials selection, coating technology, assembly and handling, and packaging methods. Many of the oft-publicized advantages of polymer optics (low cost, low density, ease of asoheric Production, etc.) can be inadvertently offset by factors which may be anticipated and avoided by a designer reasonably familiar with polymer properties, processing procedures and limitations, and their implications. It is the purpose of this paper to provide some of this information.

Thin film epoxy replication is idealy suited for the production of high quality optics. Flat surfaces to 1/10 wave accuracy are economically replicated onto metal substrates of aluminum for minimum cost and on beryllium for minimum inertia. Surface preparation is critical to obtaining good replicas. Aspheric optics are producedon both glass and metal substrates; glass being sutable for optics that do not deviate substantially from a sphere and metal being preferred for fast or difficult aspherics. A generalized aspheric telescope replicated on glass yielding near diffraction limited performance is described as an example of one of the more difficult replications.

Methods of figuring fused fiber optic arrays to achieve optical image transforms unique to the medium are discussed. A number of concepts including image plane modification, linear and non-linear magnification, and image duplicator/combiners are described. Fused fiber optic arrays have been tapered and figured, using conventional grinding and polishing techniques that will magnify an image in approximately a distance of 1.5 times the major diameter without traditional spherical or chromatic aberration considerations. It has been observed that when a parallel fiber optic array is cut at an angle other than 90° to the optical axis, an elongation of the surface, and the resulting image plane, is realized. One dimensional magnifiers up to 10X have been constructed. An element was constructed that resulted in non-linear magnification from center to edge without the necessity of aspheric generation or polishing. Triplet construction con-taining one convex/convex element and two plano/concave elements utilized spherical curves exclusively. Magnification variations of 8X have been observed.

A polishing model and a fixed charge lap model are presented. Optical polishing data is presented for 11 metals partially supporting the above.

The spatial frequency response of elastic backed flexible lapping belts is examined. The transition from planing off to propagating sinusoidal ripple is shown to be extremely sensitive to ripple frequency.

A cost effective method for fabricating lenses for infrared optical systems has been developed. The method involves press forging of halide single crystals to produce a finished lens, without need for polishing. In particular we have used the press forging approach to fabricate a KBr color corrector lens. The lens is plano-concave and was designed for use in a forward looking infrared (FLIR) imager module. A KBr single crystal is deformed by forging using a two step process in helium atmosphere at 4000 psi. The first step deforms the crystal 60% and is done at 250°C. This forging is shaped and water polished. The second forging produces a final optical surface finish and figure. The finished optical surfaces are replicated from pyrex dies at 225°C. The finished lens is optically and mechanically centered prior to testing. The surface figures of the optical surfaces are complicated but a double pass transmission test shows the wavefront distortion to be typically about 1 λ at 0.6328pm. Two wavelength holography and a shearing plate optical test both predict acceptable performance at 8-12μm.*

Optical mirrors made of beryllium can provide advantages over those made of other materials primarily because of their high elastic modulus and low density. Conventional techniques for producing mirrors can he used with beryllium but only with difficulty and the resulting surface often contains defects from the metal itself. Preliminary experiments have shown that optical mirror surfaces on metal can be replicated with vapor deposited beryllium under carefully selected and controlled conditions.

A reliable, low-cost method of bonding a thin glass substrate to an alnico VI B magnet has been developed to allow the fabrication of a high-quality optical surface while minimizing the mass of non-magnetic material. The glass is bonded with an organic adhesive. A series of heat treatments after curing insures a stable figure for substrates thinner than 0.5 millimeter.

Spherical metal mirrors have been fabricated by grinding and polishing the desired curvature in electroless nickel deposited over copper electroplated directly on the surfaces of Alnico magnets. Several different electroless nickel plating solutions were investigated, along with a variety of masking materials. Through careful selection of materials and refinements in technique a satisfactory fabrication procedure was developed.

While precision machining has been applied to the manufacture of optical components for a considerable period, the process has, in general, had its thinking restricted to producing only the accurate shapes required. The purpose of this paper is to show how optical components must be considered from an optical (functional) point of view and that the manufacturing process must be selected on that basis. To fill out this perspective, simplistic examples of how optical components are specified with respect to form and finish are given, a comparison between optical polishing and precision machining is made, and some thoughts on which technique should be selected for a specific application are presented. A short discussion of future trends related to accuracy, materials, and tools is included.

The two-axis precision diamond machining process in addition to inherent machine accuracies, relies on critical features of the diamond cutting tool and its utilization. The application of this process to IR refractive elements necessitates the extension of these implied accuracies to a higher degree than for typical reflective optics. The system requirements for a typical FLIR system are extended to arrive at typical parameters for FLIR optical elements. Their implications for the machining process in terms of such parameters as tool radial accuracies, tool centration and fixturing approach are presented. Examples of refractive elements machined at E00 and their accuracies are presented, as well as on-going developments to more closely suit the Process to the demands of large quantity production of refractive elements.

The use of a CNC 2-axis diamond turning machine has enabled a variety of aspheric optics to be fabricated. These optics, which are both reflective and refractive elements, require a testing method equally versatile. Computer-generated holographic (CGH) interferometry is a method ideally suited to this unique, emerging optical fabrication technique. The same mathematical formula used to fabricate the part is used to also design the hologram. The hologram serves as a reference master and is designed for each individual asphere. This method permits even generalized aspheres to be tested with relative ease. Many aspheres designed for use in infrared systems may be too extreme to be adequately represented by a hologram. This limitation is economically overcome through the use of simple null lenses. These lenses are used to null the lower order aberrations, thus allowing the hologram to cancel the higher order departures. The limitations of the hologram, the use and design of null lenses, and examples of testing aspheres are given.

Ultraprecision machine tools are used at the Los Alamos National Laboratory for single-point diamond turning of optics and other precision parts. Measurements of a 50-mm-diam copper flat illustrate the quality of a part that can be machined on the Moore No. 3 lathe. Measurements of a 0.4-m-diam aluminum mirror with a 20-m radius of curvature are presented as an example of a part machined on the Moore No. 5 lathe. A varying frequency sine wave grating shows a type of special optical grating that can be produced using the Pneumo lathe.

The manufacture of optical components at Ealing by the diamond machining process has reached a stage of refinement where in-process examination of surface figure has become mandatory. Currently, Ealing employs the use of interferometry to determine the surface figure of spherical, aspheric, or flat optics during the diamond machining process.

The single point diamond turning process is ideally suited to making aspheric optical elements for use in the infrared region. In this paper, applications of diamond turning to various materials and geometries for both refractive and reflective optical elements are considered, based upon direct manufacturing experience.

During the past ten years, the combined efforts of government supported organizations and a select few American industries have developed and put on line one of the most important manufacturing tools of the century - single point diamond machining of optical and optically related components. There are categories of optical elements currently being made that were heretofore too difficult to consider, too impractical or just economically unfeasible. The impact of diamond machining on common categories of optics is summarized in Table 1. In addition to these general categories, diamond machining can contribute heavily to the feasibility of producing high aspect ratio rectangular optics or thin optics that are too flexible to withstand pressures of conventional lapping and polishing.

The Lawrence Livermore National Laboratory (LLNL) Machine Tool Development and Machine Control Groups have recently completed the construction of a small, high-accuracy, two-axis numerically controlled Diamond Turning Machine. The machine, which was assembled from commercially available components, cut its first contoured part two weeks after the project was initiated, and was placed in service after a total of three weeks. One-inch diameter hemispherical aluminum test parts have been cut to an accuracy of 10 microinch (size and contour band) with a surface finish of 1-1.5 microinch rms. The machine, which is limited to parts less than four inches in diameter, has been designated the Baby Optics Diamond Turning Machine (BODTM).

Although computer holograms have been discussed often for optical testing, practical constraints on the size of hologram which can be written have prevented widespread use. We report here development of hardware and software to create holograms with fringes as fine as 100 lines /mm with fringe centers absolutely located to 0.25 pm over a diameter of 85 μm. This gives the equivalent of 3.3 x 10 5 points across a diameter.

Cryogenic application of mirrors for laboratory purposes and in spaceborne instruments has caused a rapid increase in the use of metal mirrors in the field of optical engineering. The recently constructed Lockheed Sensor Test Facility (LSTF) uses metal mirrors in cryogenic applications. This article discusses some of the issues involved in the design and fabrication of the LSTF mirrors, such as the problems of material selection, design of mirror shape, and the heat treating process. The final LSTF mirror design and fabrication process are described, as well as the use of the mirrors in the Test Facility and the results of tests performed.

The use of computer-generated holograms (CGHs) to test aspheric surfaces fabricated by modern optical methods such as diamond-turned machining has become increasingly important. The making of CGHs may, however, be limited in spatial resolution and space-bandwidth product provided by commercial optical recording devices. We will demonstrate that CGHs of high spatial resolution and large space-bandwidth product can be written directly on electron-resist using e-beam lithography. This approach not only reduces plotti9g errors normally introduced by optical recording devices, but also provides more than 100 distor-tion-free resolution picture elements in a synthetic hologram of correct size. In this paper, we will discuss how to make such a synthetic hologram by means of e-beam lithography. The performance of this CGH will be demonstrated by comparing with a non-rotationally symmetric aspheric wavefront of over 100 waves of spherical aberration using a concave mirror and plane parallel plate combination as the test piece.

An optical component test system was designed and fabricated. The system was optimized for automation, speed and component flexibility. The optical component or system is characterized according to parametric measurements including focal length, optical transfer function, scatter, and spectral characteristics. The OTF processing is performed by an analog technique which provides increased speed and frequency selection capability over conventional digital systems.

The Space Telescope is a 2.4-meter aperture telescope coupled to five on-board astronomical scientific packages. The accurate line-of-sight stability required is provided by the Pointing Control System, whose optical system is the Fine Guidance Subsystem (FGS). The collimating mirror within the FGS optical system is aspherized in order to correct the pupil aberration. A null corrector is required in order to test the collimating mirror in autocollimation. Two alternate null corrector designs and their tolerance analyses are described and a preferred design is selected for fabrication.

A system has been developed for the purpose of real-time rapid measurement of the optical path difference (OPD) between a reference wavefront and a measuring wavefront of an interferometer by measuring the phase difference between them. The system is capable of measuring accurately OPDs represented by interference patterns with any shape and degree of complexity within the spatial resolution limits of the detector. The interference pattern of the two wavefronts is modulated so that any given point in the interference plane has a sinusoidally varying light intensity with a phase that is proportional to the OPD between the two wavefronts. The individual diodes of an array camera detect the varying light intensity in the interference plane. The output voltage of each diode is digitized and stored. The phase is then computed from these data, and processed to obtain the OPD, peak-to-valley (P-V), and root-mean-square (rms) values of the measured wavefront.

A new process at Xerox Electro-Optical Systems, which stems from an interferometer equipped with a cylindrical transmission lens, is the ability to in-process test cylinders on-the-block. This capability permits evaluation of the surface at the most economical point, while preserving the capability to correct same. Furthermore, expensive cylindrical test plates, which can each run hundreds or thousands of dollars, can be eliminated. In addition, startup time or change time is reduced, since the potentially long wait for fabrication of cylindrical test plates is eliminated.