AN IN UNITS OF

10

be obtained by repeated testing of mirrors for the Orbiting Astronomical Observatory (OAO) and Apollo Telescope Mount (ATM) programs.

The situation is even less satisfactory with the other materials. Stable beryllium mirrors can be manufactured in sizes up to approximately 20 inches (0.5 meter), and silicon samples up to 12 inches (0.3 meter) have been prepared. Very extensive development is required before beryllium and/or silicon mirrors 1 to 3 meters in diameter can be manufactured. The earlier remarks about testing would also apply to these mirrors.

Beryllium oxide has a favorable combination of physical properties for mirror applications. Unfortunately, commercial manufacturing practice yields low density (96 percent) and hence poor reflectivity. Translucent beryllium oxide has been produced in the laboratory in small samples. A very considerable amount of binder is added to the translucent material; this may detract from the dimensional stability.

1.0

Materials Technology Development Required Prior to Orbiting a Large Astronomical Telescope

At this point in time, it is tempting to assume that there will be several orbiting observatories, each with its own prime mission, such as solar observation, survey of the universe, etc. With this philosophy, we can say with safety that optical figure and dimensional stability will be the primary technical objectives of the designers, with weight and thermal response alternating between second and third in priority.

Because of the massive development required to perfect the manufacturing process and control for any material, the number of candidates should be limited as rapidly as possible. At this time, I feel that we have four candidates: beryllium, silicon, Cer-Vit, and ULE. In some ways, this is a very uneven competition. Although the military and space programs have prompted the rapid development of beryllium technology, new facilities will be required to manufacture large mirrors.

The semiconductor application of silicon has made it one of the best understood materials known to man. Again, manufacturing techniques must be developed before large mirrors can be prepared. The manufacturing technology of the new dielectric materials is in much better shape. On the other hand, the "unavailable" materials appear to be superior to the new dielectrics. This impasse can be broken by the following steps:

1. Measure the effect of ionization radiation on the dimensional stability of small samples of all four materials.

2. Extend the dimensional stability measurements to cryogenic temperatures.

3. After a period of scale-up, measure the dimensional stability of several fairly large mirrors of each material that passed the radiation test.

Step 3 is time-consuming and expensive, but the penalty for not employing the best material is incalculable. If the mirror is unnecessarily heavy, some instruments will be omitted; and some astronomical data will remain uncollected. Conversely, if the figure

of the mirror changes drastically within a short period of time, the telescope will become more or less useless.

Mirrors in the 1-meter-diameter range could be manufactured from each material, mounted on a standard mount, tested, employed by astronomers in the mount, and retested at intervals. This sequence is suggested because the mirrors will be of better quality than terrestial observatories can utilize and because test facilities are expensive.

A parallel effort during this time period should be the optimization of the design for the mirror and mount. A continuing dialogue between the numerous technical disciplines is essential throughout this development phase. Such a dialogue is justifiable for, if the system planners elect to use a cold mirror without informing the materials people, the radiation and stability measurements will be made incorrectly.

With careful planning and development, I feel that large, dimensionally stable, diffraction-limited optics can be placed in

orbit.

Introduction

More mirrors over 100 inches in diameter are currently being fabricated or planned than have been completed to date. As a result, there is a shortage of experienced personnel available to fabricate these pieces, which must necessarily be worked by people whose experience has been limited to smaller sizes. Difficulties can occur when tools and techniques that have been proven for smaller mirrors, 40 to 60 inches in diameter, are used to fabricate large mirrors. Therefore, I will discuss some of the problems we at Fecker Systems Division have encountered in this area and some of our solutions to these problems. I would be less than honest if I claimed these are the right answers, for optical fabrication techniques vary greatly from optical shop to optical shop and between individuals in the same shop. Furthermore, I am limited to the problems that we have actually encountered and to the solutions and rules of thumb that have worked for us.

Microripple

The first problem that occurs in working large mirrors is the decrease in the size of the allowable slope errors. A 4-inch-diameter mirror can resolve 1 arc second, and slope of 1/4 of an arc second should not detract appreciably from its performance. A 100-inch diameter mirror can resolve 1/25 of an arc second, and slope errors must be limited to 1/100 of an arc second to utilize it to its fullest. Thus, the larger the mirror (everything else being equal), the more carefully it must

be fabricated and tested. Not only must zonal irregularities be virtually eliminated, but also the "texture" or microripple produced in the mirror's surface by the polisher must be controlled. This microripple may be virtually invisible to both knife-edge and interferometric testing, but it causes light to fall outside the predicted blur circle. The appearance and degree of this microripple depends on polishing pitch hardness, polishing compound, size and configuration of the facets in the polishing surface, polishing pressure, velocity of the polishing tool relative to the workpiece, the speed and asphericity of the mirror, and the size of the polishing tool relative to the mirror. Although very little work has been done concerning the control of microripple, it appears that large polishers, large but unequally sized facets, soft polishing pitch, light polishing pressure, and slow polishing speed tend to minimize or eliminate microripple. The degree to which any measures may be pursued is limited by other considerations. For example, polishing pitch soft enough to completely eliminate microripple may cause an unacceptable "roll-off" at the edge of the mirror; extremely large facets may cause an unacceptable increase in "coldpressing" periods. If, however, care is taken, techniques can be found that suppress microripple while allowing the mirror to be figured.

Compliance

The compliance or lack of stiffness of large mirrors causes many difficulties unless carefully considered. The compliance of a mirror varies as the square of the diameter for a fixed diameter-to-thickness ratio. A

100-inch-diameter mirror that is 12 inches thick is equivalent to a 40-inch diameter mirror approximately 2 inches thick. Most opticians would consider the larger, thicker mirror less demanding and the thinner, smaller mirror more of a challenge although the two are equivalent. The problem of the compliance of larger mirrors can be solved easily by a well-designed conservative support system. We use a relatively simple relationship for determining the number of supports required. Although the derivation of this relationship is not rigorous, it has proved to be satisfactory for a variety of pieces, ranging from 40 inches in diameter by 4 inches thick, to a mirror 80 inches in diameter by 3 inches thick, to a mirror 103 inches in diameter by 12 inches thick. The number of support points is:

It should be noted that these are the number of points necessary to support only the mirror itself, not the mirror and polishing tools. The number of support points must be increased depending upon the weight and stiffness of the polishing tools.

A great variety of systems have been successfully used to support mirrors, including carpeting, foam rubber, and even kinematic devices. Large mirrors deserve supports more sophisticated than foam rubber, but kinematic systems become cumbersome. Figures 1 and 2 show a support system used extensively at Fecker for a variety of work pieces. Each support is a mechanical spring that has been individually calibrated and cut to length to ensure uniformity from point to point. Each spring is permanently attached to the table of the polishing machine.

To prevent the mirror from rocking as the polisher moves over the mirror, some (or

Figure 1. Uncovered view of individual pitch-spring.

pected and somewhat unpredictable manner. The effects produced by such a tool are similarly unexpected and unpredictable. In addition, a compliant tool will tend to conform to the surface of the mirror and will not remove cylinder or astigmatism. If a large tool is designed so as to maintain stiffness, it must necessarily become heavy and thus further increases the pressure between the polishing tool and the mirror. There seems to be little choice in designing, polishing, or grinding tools for a large mirror; if they are to work in a predictable manner and are to produce a regular surface, they must be stiff. The resulting increase in polishing pressure, however, must be compensated. This has been done most frequently by counterbalancing the tool. An alternate approach has been used successfully at Fecker. This is a stiff, lightweight tool fabricated from two plates and a honeycomb core. These three pieces are joined together by epoxy. The resulting tool can be sufficiently stiff to be predictable yet light enough to need no counterbalancing. We have used such tools up to 100 inches in diameter. Such a tool is shown in figures 4 and 5.

Difficulties also occur in the reduction of a high intermediate zone on a large mirror. With small mirrors, the optician may simply alter the shape of the polishing tool, scraping away the pitch from the areas corresponding to low zones and leaving a high area corresponding to the high area to be removed.

Figure 3. Cross-section of pitch-spring support.

Grinding and Polishing Tools

The compliance of both grinding and polishing tools varies in much the same way as the mirror itself. If a simple ribbed casting is scaled, the compliance varies as the square of the diameter and the unit pressure exerted upon the workpiece by the polisher varies directly with diameter. Thus, a large, compliant, polishing tool will flex and deform when overhanging the mirror in an unex

Figure 4. Honeycomb tool, used without counterbalancing.