In the second half of the 1990's, LLNL and others will be designing and beginning construction of the National Ignition Facility (NIF). At more than 10 times the power and size of the Nova laser system, this new laser will be capable of producing the worlds first controlled fusion ignition and burn, completing a vital milestone on the path to fusion energy. In order to optimize the performance of the laser system for a minimum cost, we have been conducting a campaign to properly specify the optical properties of the more than 7,500 large optical components to be deployed in the NIF. The draft optics specifications derived from this effort will be presented. The evolution of these specifications, both in language and in content, will be discussed, specifically transmitted wavefront (both P-V and PSD), scratch/dig, surface roughness, bubbles and inclusions specifications.

This paper discusses the results of high spatial resolution measurement of the transmitted or reflected wavefront of optical components using phase shifting interferometry with a wavelength of 6328 angstrom. The optical components studied range in size from approximately 50 mm X 100 mm to 400 mm X 750 mm. Wavefront data, in the form of 3D phase maps, have been obtained for three regimes of scale length: 'micro roughness', 'mid-spatial scale', and 'optical figure/curvature'. Repetitive wavefront structure has been observed with scale lengths from 10 mm to 100 mm. The amplitude of this structure is typically 1/100 to 1/20. Previously unobserved structure has been detected in optical materials and on the surfaces of components. We are using this data to assist in optimizing laser system design, to qualify optical components and fabrication processes under study in our component development program.

This paper describes the use of Fourier techniques to characterize the wavefront of optical components, specifically, the use of the power spectral density (PSD) function. The PSDs of several precision optical components will be shown. Many of the optical components of interest to us have square, rectangular or irregularly shaped apertures with major dimensions up to 800 mm. The wavefronts of components with noncircular apertures cannot be analyzed with Zernicke polynomials since these functions are an orthogonal set for circular apertures only. Furthermore, Zernicke analysis is limited to treating low frequency wavefront aberrations; mid-spatial scale and high frequency error are expressed only as 'residuals'. A more complete and powerful representation of the optical wavefront can be obtained by Fourier analysis in 1 or 2 dimensions. The PSD is obtained from the amplitude of frequency components present in the Fourier spectrum. The PSD corresponds to the scattered intensity as a function of scattering angle in the wavefront and can be used to describe the intensity distribution at focus. The shape of a resultant wavefront or the focal spot of a complex multi- component laser system can be calculated and optimized using the PSDs of individual optical components which comprise it.

LLNL plans to specify optical components for the National Ignition Facility according to ISO 10110, the new international standard for preparation of optics drawings. The standards have been approved by the international optics community and represent a fairly comprehensive language for describing optical components. We will describe our plan for implementation and experience to date in doing so.

Optical image differentiation operations by means of the wavelet transform (WT) are given in this paper. We detail the advantages of Haar's function and its various types as differentiation wavelets over other functions. We give the appropriate type and parameters of the functions they can demonstrate the best effects of image differentiation. In this paper two types of power spectrum nonlinear processing of joint Fourier transform correlator in hybrid system are also reported. Both the processing of fractional power of Fourier spectrum and the filtering of the zero order spectrum have some special properties, respectively. Combined the Haar's wavelet transform with the power spectrum nonlinear processing, we demonstrate the differentiation of the upper right or down left of optical iamge.

The use of a fiber optic reference to test cylindrical optics in a Fizeau interferometer configuration is described. (Note: this presentation was accompanied by a short video demonstration).

Two methods of calibrating an interferometer for testing cylindrical surfaces are tested experimentally. One method uses a tilting mirror instead of the test specimen and combines at least two measurements at different angles. The other method adds together three measurements. In two of them a cylindrical specifmen is used. In the third one the specimen is replaced by a roof mirror. Results of both methods are computed. They show a reproducibility of (lambda) /50 and an accuracy of (lambda) /20 for the absolute surface deviations of a cylindrical specimen.

Results from a new technique for the absolute calibration of flats are presented. This technique is mathematically exact and, in addition, computationally nonintensive compared to previous techniques. Hence it allows the calibration of flats to a high degree of precision over the entire surface. The accuracy of this new technique is demonstrated through simulated test cases. This technique is currently implemented in the manufacturing cycle of optics (RMS precision of approximately (lambda) /250 at a wavelength of .6328 microns).

Ion figuring is an efficient and deterministic process for fabricating high quality optical surfaces. The simultaneous phase shifting interferometer is a helium-neon laser-based Twyman-Green system capable of data acquisition at shutter speeds of up to 1/10,000th of a second, which eliminates vibration effects. Ion figuring and simultaneous phase shifting interferometry were successfully combined to produce an aspheric, 18 inch diameter, f/2, ULE mirror of exceptional surface figure quality. Through iterative cycles of SPSI testing and ion figuring, the mirror surface was processed to a final figure quality of 0.007 (lambda) rms (95% of data <EQ 0.028 (lambda) p-v). The radius of curvature of the mirror was held to within 0.00008 inch of the design radius. Details of the testing/figuring cycles are discussed along with information regarding the mirror mounting scheme, interferometer calibration, and data analysis techniques.

When characterizing an extreme ultraviolet (EUV) lithographic optical system, visible light interferometry is limited to measuring wavefront aberration caused by surface figure error while failing to measure wavefront errors induced by the multilayer coatings. This fact has generated interest in developing interferometry at an EUV camera's operational wavelength (at wavelength testing), which is typically around 13 nm. While a laser plasma source (LPS) is being developed as a lithography production source, it has generally been considered that only an undulator located at a synchrotron facility can provide the necessary laser-like point source for EUV interferometry. Although an undulator-based approach has been successfully demonstrated, it would be advantageous to test a camera in its operational configuration. We are developing the latter approach by utilizing extended source size schemes to provide usable flux throughput. A slit or a grating mounted in front of the source can provide the necessary spatial coherence for Ronchi interferometry. The usable source size is limited only by the well-corrected field of view of the camera under test. The developemnt of this interferometer will be presented.

We present a simple and general method of aspheric figure metrology using a CGH null mounted in the test beam of a conventional Fizeau or Twyman-Green interferometer. A 'standard' reflective CGH is used to establish optical alignment with respect to the interferometer's spherical test beam. This alignment is then mechanically trnaferred to a custom CGH null. The accuaracy of the alignment transfer is readily verified. The test method has been modeled by raytracing and verified experimentally by testing a perforated 8 inch F/1.5 on-axis paraboloid and a 50 mm off-axis paraboloid from their centers of curvature.

Computer generated holograms (CGHs) are an alternative to refractive or reflective null optics when testing spheric optic components. A key attraction is that the difficulty of designing and fabricating a CGH null is largely independent of the detailed shape of the test asphere. CGH nulls have been used quite successfully in a number of high profile programs, but certification issues have limited their more widespread acceptance as a primary testing means. This is due largely to unfamiliarity with appropriate verification and certification methods. We here discuss specification and tolerancing of CGH nulls and present a comprehensive methodology for verification and certification.

Measurements of large convex surfaces are notoriously difficult because they require auxiliary optics that are larger than the surface being tested. Fizeau interferometry is well suited for these surfaces because the only surface required to be made to high accuracy is the concave reference surface, which is only slightly larger than the surface being measured. Convex surfaces which are spherical or aspheric can be measured using spherical, aspherical, or holographic test plates. The reference surfaces for these tests must be of good quality and measured to high accuracy. The optical systems that provide illumination and create an image of the interference pattern do not have to be made to high quality. The illumination systems can typically have errors several orders of magnitude larger than the allowable surface measurement errors, so these systems can be made at low cost. Several such systems using low cost aspheric mirrors and lenses for measuring convex spherical and aspherical surfaces are presented.

Aspheric surface testing would be greatly facilitated if the requirement for a null condition were removed. Testing an optic in a non-null configuration introduces aberrations into the wavefront. The wavefront measured at the sensor is different from the wavefront initially produced by the test surface, and the interferometer must be calibrated if useful measurements of aspheres are to be made. One potential calibration technique is reverse optimization, where a lens design program is used to retrieve the prescription of the interferometer. Various problems in modeling an interferometer, and potential solutions, are discussed. A defocused sphere was used to generate a non-null wavefront with 100(lambda) of departure at the surface. The reverse optimization results matched the experimental data to better than (lambda) /4 PV.

In the methods to test an asperhical optical surface, sheering interferometer is a testing instrument by which the wavefront of the surface testing is split apart into two and then interference fringes can be formed without demanding a standard reference aspherical wavefront. However, if the sheering interferometer is applied to test an aspehrical surface, having a relatively big deviation from its best fit spherical surface, the fundamental measuring error will increase very quickly. A comprehensive analysis of this error is given and an improved method is put forward by which the testing scale of the sheering interferometer is enlarged by introducing an analogue standard wavefront produced by a computer so as to reduce error resources and to improve the testing accuracy.

The usage of two-wavelength holographic interferometry (TWHI) in testing aspherics and evaluating the testing results are discussed in the paper. TWHI accomplishes actual long- wavelength interferometry, using visible light as working wavelength. The sensitivity, accuracy, and range of measurement can be adjusted conveniently.

Axial intensity scans provide a noninterferometric method for measuring aberration content of a focused beam. Applications of the technique include the determination of the conic constant of aspheric mirrors (without the need for a null lens).

A profilometer is described that utilizes the swing-arm geometry to provide surface profile measurements of large, highly aspheric surfaces. The profilometer measurement is shown to be robust against stiffness and alignment induced errors in the probe motion.

To measure profiles of an object with 3D free-form surfaces, a new optical noncontact sensor is proposed in this paper. Two linearly polarized laser beams from the sensor, which are used for rough and fine positioning of the sensor, are focused on an object surface to be measured. The depolarized components of backscattered light from the object surface are detected, with the exception of specularly reflected light, using optical components. The triangulation method is applied to the sensor in order to obtain a signal which is proportional to the displacement of the object surface for rough positioning of the sensor. Moreover, the astigmatic focus error method is applied to the sensor in order to detect a focused positioning signal of the object surface for fine positioning of the sensor. The measuring instrument for 3D profiles of an object surface mainly consists of a newly designed optical noncontact sensor and tables. In the case of detection based on the astigmatic focus error method for fine positioning of the sensor, one special light detector for the sensor is proposed. Using the measuring instrument, 3D profiles of diffuse reflection surfaces and metal surfaces on which both specular reflection and diffuse reflection occur can be measured. Adverse effects of surface roughness, scratches, and diffraction on measurements can be reduced.

The use of aspheric and unconventional optical components has rapidly increased over the past few years. But the fabrication techniques of such components is limited by lack of suitable methods for assessment of surface shape of large aspheric optical components, especially for large convex surface. In this paper we discuss a 3D pencil beam profiler which is suitable for measuring optical aspheric surfaces, having different parameters and a large radius with interferometric accuracy without the need of standard surface an null lens. The principle of the profiler is based on a pencil beam optical scanner. The tested workpiece is put on a turnable table with a standard plane plate which is used to give the information of the tilt of the turbable table when the table is rotated for 3D contour measurement. The information will be applied for removing the error introduced during the table turning from the final map of the surface. The measurements are made along the meridian of the tested surface. The input slope data of the meridian are converted to Taylor series by least square fitting and to get the height map of the surface. After repeating above procedure on several lines on the tested surface, Zernick polynomials are employed to build up 3D map of the tested surface according to the data from the Taylor series and azimuth information of the turnable table. Examples of using this profiler to test optical components are given in the paper. The outcome is compared with the result given by the interferometer and the difference between two results is less than 0.05 wavelength.

An overview of successful optical apprenticeship programs in the Rochester area where both academia and industry have created partnerships to develop and deliver industrial training material together.

Several aspheric lens elements have recently been produced using new manufacturing processes. Deterministic microgrinding and magnetorheological finishing technologies were employed to fabricate ten 47 mm diameter plano convex hyperboloidal aspheric condenser lenses. The parts had an aspheric departure of 145 microns from best fit sphere, with a final figure requirement of approximately one micron p-v. The lens elements were tested at the component and systems level and passed all end user requirements.

This paper describes an online noncontact metrology system for tool misalignment detection on the Opticam spherical surface generator. The system detects distortion due to surface errors in the reflected image of a bar target. The amount of image distortion is determined by the relative slope changes on the test surface, and it can be calculated as a phase function of the reflected image. The phase function can be used to determine surface slope variation of the test part, and this slope information can be used to determine the cutting tool misalignment in the Opticam system.

For precise deterministic microgrinding of optical components, the exact location of the region of contact between the machine tool and the glass work piece needs to be specified. Any tool compliance therefore contributes to a lack of precision during grinding and to error in the final surface of the glass work piece. The goals of this work are to analyze the dependence of ring tool compliance changes and to develop simple analytical formulas that express these relationships for machine tools used in spherical lens fabrication. Numerical predictions have also been obtained using NASTRAN, a linear finite element analysis routine. The numerical results have been compared with estimations obtained from approximate analytical expressions to determine the ranges of validity of the expressions. These simplified expressions can be used to aid the design of ring tools for smaller [O(mm) diameter] and larger [O(m) diameter] optical lens fabrication.

LLNL is collaborating with the Center for Optics Manufacturing and the American Precision Optics Manufacturers Association to optimize bound diamond ring tools for the spherical generation of high quality optical surfaces. An important element of this work is establishing an experimentally-verified link between tooling properties and workpiece quality indicators such as roughness, subsurface damage, and removal rate. In this paper, we report on a standardized methodology for assessing ring tool performance and its preliminary application to a set of commercially available wheels. Our goals are to 1) assist optics manufacturers (users of the ring tools) in evaluating tools and in assessing their applicability for a given operation, and 2) provide performance feedback to wheel manufacturers to help optimize tooling for the optics industry. Our paper includes measurements of wheel performance for three 2-4 micron diamond bronze-bond wheels that were supplied by different manufacturers to nominally-identical specifications. Preliminary data suggests that the difference in performance levels among the wheels were small.

Textured laps can be created by 'slumping' uniform thin films of appropriate materials over textured substrates that have been generated to the required figure. The film can easily be replaced and the lap shape is invariant since the lap substrate never contacts the workpiece nor the polishing compound. Thus extremely repeatable lapping and polishing processes are obtained. This paper introduces this approach to lap construction and presents preliminary results.

Because there are no practical commercially available bound abrasive polishing media, we are developing a bound abrasive polisher for deterministic finishing of optical glasses. Several in- house formulated polishing pellets, molded laps, and ring tools have been studied. Two experimental test beds were employed. The first involved the polishing of flat optical glass parts on single pellet and molded pellet laps. The tests were conducted on a single spindle machine. The performance of in-house manufactured laps was compared to experimental and commercial formulations obtained from industry. Compositions which polished the glass below 20 angstrom rms surface roughness were selected for additional testing. The second test bed for these formulations was the Opticam SM. Materials were molded into a ring tool geometry. Although the tools polished effectively, more work is required to control surface figure during final finishing.

The problem of a subsurface crack parallel and close to the surface of a semi-infinite medium was studied by dislocation modeling and finite element analysis. The loading is applied over the surface of the semi-infinite medium. For tensile loading the dislocation model gives the same result as the finite element method. For shear loading, the crack faces penetrate each other for the traction free crack surfaces. Using the ABAQUS code and the 'interface' or 'gap' elements over the crack faces, the overlap problem was avoided. It is found that one end of the crack is closed and the other is open and the Mode II stress intensity factor at the closed crack tip is larger than that at the open tip. All the stress instensity factors increase as the subsurface crack approaches the surface.

We discuss a constitutive model describing the permanent densification of fused silica under large applied pressures and shear stresses. The constitutive law is assumed to be rate- independent, and uses a yield function coupling hydrostatic pressure and shear stress, a flow rule describing the evolution of permanent strains after initial densification, and a hardening rule describing the dependence of the incremental densification on the levels of applied stresses. The constitutive law accounts for multiaxial states of stress, since during polishing and grinding operations complex stress states occur in a thin surface layer due to the action of abrasive particles. Due to frictional and other abrasive forces, large shear stresses are present near the surface during manufacturing. We apply the constitutive law in estimating the extent of the densified layer during the mechanical interaction of an abrasive grain and a flat surface.

High quality optics have been manufactured using both conventional spindle polishing and small tool figuring under computer control. While both techniques have been used to produce precision optics in the past, this paper describes the hybrid process of using both processes to utilize their specific figuring efficiencies. By knowing the frequency content of the surface being corrected, the proper tool size can be selected to address the dominant frequency. By using the most efficient tool for correction fabrication time will be reduced yielding more cost effective processes. Proper process calibrations and controls are required to maximize convergence and material removal.

In this paper, edge control algorithm is studied based on a modified removal function, a computer simulation program is written to provide information on edge figuring process. The reliability of the algorithm is confirmed by CCP polishing experiment. The experimental results appear to have good agreement with theoretical analysis, spherical surface with edge errors of 2.56(lambda) P-V ((lambda) equals 6328 angstrom) is polished for 1.5h, and the errors reduce to 0.89(lambda) P-V.

We describe the optical grinding and polishing, and the measurement of sphericity by multiple roundness profiles, of large single crystal silicon spheres (balls) (93.6 mm diameter and 1 kg mass). The balls are used in direct measurements of volume and density in the determination of the Avogadro constant. A typical peak to valley error of sphericity achieved was < 60 nm. The arithmetic mean deviation of the error surface relative to a least square sphere was around 16 nm. The microroughness was < 0.2 nm rms when measured with a WYKO TOPO 3D noncontact surface profiler. Defects in the optical finish were negligible when compared to the total allowable uncertainty specified for the volume of the balls of 0.1 ppm.

Intercomparison roughness measurements have been carried out on supersmooth artifacts fabricated from BK7, fused silica, and Zerodur. The surface parameters were determined using the optical heterodyne profiler Z5500 (Zygo), a special prototype of the mechanical profiler Nanostep (Rank Taylor Hobson), and an Atomic Force Microscope (Park Scientific Instruments) with an improved acquisition technique. The intercomparison was performed after the range of collected spatial wavelengths for each instrument was adjusted using digital filtering techniques. It is demonstrated for different roughness ranges that the applied superpolishing techniques yield supersmooth artifacts which can be used for more intercomparisons. More than 100 samples were investigated. Criteria were developed to select artifacts from the sample stock.

The prevailing technique for assessing surface roughness and related parameters from engineering surfaces in industry is mechanical stylus profilometry. Though it is a conceptually direct method it suffers from several drawbacks; the measurement is not area-covering while many important features of engineering surfaces are related to 3D topography; the technique is slow; and, the technique involves contact with the surface, potentially damaging it. For surfaces of optical quality, elastic light scattering provides an efficient way of characterizing surfaces, and two ASTM standards exist pertaining to such measurements. However, the theoretical models based on first order perturbation theory break down when the rms roughness approaches one tenth of the incident wavelength. We have used infrared scattering at 10.6 micrometers wavelength to provide area-covering, non-contact surface characterization of engineering surfaces. The surface power spectral density is calculated from the scattering distribution by Rayleigh-Rice vector perturbation theory. Measurement of bandwidth limited surface roughness is demonstrated for ground and polished surfaces in the root mean square roughness range of 0.03 micrometers to 1.7 micrometers . Good correlation with contact stylus measurements is achieved for nearly 1D surfaces, anisotropic surfaces, and isotropic surfaces. Measurement of scratch and lay characteristics are also demonstrated.

An optoelectronic device for automatically inspecting optical surface defects by recognition of forescattering patterns is described. A light spot coming from a diode laser scans the optical surface following a track of spiral. The normal light falls into an optical trap. The forescattered light by the surface defect is projected on a CCD array where the forescattering patterns are formed. Different kinds of defects make very different patterns. By recognition of the patterns using a computer, the shape and size of a defect are obtained simultaneously. The device can inspect the defects both on the surface and under the surface efficiently. The light spot takes the shape of a sharp line in order to improve the detectability of the scratches. Through a rotating Dove prism, this sharp line spins while spirally scanning the surface so as to meet the surface scratches. The effect of dust on the surface under examiniation is minimized. The experimental results show that this device can detect the scratches which are only just visible to a trained inspector.

The 11-meter primary mirror of the Hobby-Eberly Telescope consists of 91 hexagonal segments. Each segment is 1 meter across (flat to flat). The unique design of this telescope allows for a spherical radius of curvature on each segment. Requirements are that each segment's radius of curvature match to within +/- 0.5 mm of the nominal 26,165 mm radius, and that the surface figure be within 0.033 micrometers RMS. The optical fabrication, testing, and assembly of these segments will be discussed along with a description of the segment mounting scheme.

The fabrication and testing of a corrector mirror for a reflective Schmidt camera is to be presented. With a reflective corrector, the incident beam has to be off-axis. The corrector is tilted at an angle of 30 degrees with the optical axis to implement this requirement. The figure of the corrector, which composes of the toric term, rotational symmetrical and rotational nonsymmetrical aspherical terms, poses a certain degree of challenge for fabrication. The total correction is 46 waves peak to valley at 632.8 nm wavelength.

The precision of beautifully finished quadratic surfaced mirrors is about +/- 1/10000th approximately +/- 1/15000th of a radian, performed by the experts, they are called 'Master'. These performances 'beautifully finished' have been done by Master's intuitions or experiences, so it is difficult for mass production of a high quality product. We are trying to reach the Master's territory by using mechanical methods. These are cam and link methods, link-gear-cam method, and glass gauge method. In these methods, the large solar furnace has been produced by cam method, the mirror was ground directly, then mirror precision reached in +/- 1/6000th aprroximately +/- 1/8000th of a radian. In another method, link method, link-gear-cam method, and glass gauge method are used for production of quadratic surfaces master molds, and the glass plates are softened for the quadratic surfaced mirror by these master molds. For example, the precision of the master molds which was fabricated by the cutting machine constructed from the link-gear-cam mechanisms are +/- 1/3000th of a radian, and the precision of master molds which was fabricated by the glass gauge cutting machine are +/- 1/6000th approximately +/- 1/7000th of a radian. Lately, we have considered the quadratic cutting machine constructed from the cylinder type tool for the large diameter of the quadratic surfaces.

The High Altitude Observatory (HALO) is an instrumented gulfstream IIb aircraft sponsored by the US Army Space and Strategic Defense Command as an optical data collection platform. Capable of operation at altitudes above 50,000 feet, the HALO's highly diverse and flexible sensor suite has provided infrared, visible and ultraviolet data on numerous BMDO, DOD, and NASA test programs. This paper describes the design, fabrication, and testing of two large zinc selenide infrared windows employed by the aircraft's primary sensor system, the Infrared Instrumentation System. The paper describes the design, fabrication, and testing of the windows and coatings, the window proof testing for flight qualification, and summarizes the current operational experience with the windows.

On the basis of long-term experience of LOMO PLC in creating large optical systems for ground and space telescopes, with diameter of primary mirror from 1 to 6 meters, the following issues should be considered: principles of constructing optical systems for space telescopes and selecting their optimum design in respect of dimensions/mass and performance criteria; ensuring the fulfillment of image quality requirements in the process of manufacturing optical systems for controlling ground telescope elements in operating conditions; providing automatic adjustment of telescope secondary mirror, automatic focusing, interferometric control of image quality by means of stellar interferometer with radial shift and internal control with Gartman's test. Description of space telescope equipped with primary mirror of diameter 1.5 m, manufactured in LOMO PLC, is given.

In this report are considered the basic problems which relate to developemnt, manufacture, experimental trying out, and usage of primary mirrors (PM) of the large space telescopes intended to perform distant sounding of the Earth. Attention is concentrated on development of weight-reduced passive mirrors which ensure more reliable operation of the telescope as a whole. In the report we expressed the opinion that it is quite possible to manufacture a passive weight-reduced PM if its diameter is equal approximately to 3 m. Materials which may be used for the manufacturing of PM are beryllium and silicon carbide, physical and mechanical parameters of which are the most preferable ones. But it should be taken into consideration that this is the glass ceramic of CO115M brand which has been mastered by the industry of Russia in the greatest extent. It was confirmed that parameters of this material remain unchanged during a long period of time. Constructions of the PM, made of glass ceramic, as well as constructions of holders intended to fix the mirror, are presented in this report. A holder is used first of all to prevent lowering of a PM surface quality after a mirror has been removed from a machine and fixed in a primary mirror assembly (PMA). At present two-layer construction of a PM is preferable. This construction consists of thick base including weight reduction structure, which is in a radius which is optimum from the standpoint of deformation of a mirror operating surface. In the process of manufacture a mirror is deprived of its weight with the use of special pneumatic off-loading elements. PMA is erected in vertical plane by means of using an interferometric inspection system. In the end of this report we expressed the views on an approach to engineering of a PM by taking into account potentialities both of space ships and of carrier rockets.

The performance and utility of optical systems can be significantly improved by using lightweight stable mirror components. Such components have been incorporated into a mirror assembly through a process that utilizes low temperature fusion to bond low CTE faceplates onto a lightweight core structure. The core structure is fabricated using an abrasive waterjet cutting technique that enables the designer to optimize core geometry to enhance the structural performance of the blank. The faceplates are bonded to the lightweight core resulting in a stiff structural mirror. A 15.125 inch Zerodur mirror was fabricated using this unique process and subjected to a test program to measure the optical stability of the mirror in support of either space-based applications. Strength testing was performed to verify the integrity of the fused joints and determine appropriate design allowable stress. A thermal test program was designed to assess performance at various temperature extremes and a mechanical load test was run to verify the capability of the blank to withstand operational loads without degrading the optical surface. The surface figure of the part was measured before, during, and after a 200 degree F temperature cycle with no change in figure quality. In addition, the mirror was subjected to a 10 g load for ten minutes with no change in the figure quality.

The manufacture of 8m class ULE^TM glass telescope mirror blanks involves multiple steps each of which can induce both stresses and flaws. To ensure the mechanical reliability of such large mirror blanks, both during and post-manufacturing, extreme care is required in the various process steps (including handling, grinding, finishing, and transportation) so as to limit the stress and flaw severity to below the threshold level for the specific surface finish. This paper examines the critical stress/time histories the blanks experience during manufacturing and their ability to initiate slow crack growth from grinding flaws. Two different surface finishes, namely 120 grit and 270/325 grit, with and without acid etching are characterized with respect to strength, flaw, and fatigue behavior. These data show that the 270/325 grit finish with acid etching is appropriate for the sagging step which imposes a static stress of 650 psi in the center region of the mirror blank over a two-week sag period. Similarly, the 120 grit surface finish is adequate for grinding, packaging, and transporting steps which impose a static stress of 435 psi in the support pad region for a three month period. These predictions, based on power law fatigue model, were verified by conducting static fatigue tests (at appropriate stress levels) on a large sample of 6 inch diameter ULE^TM discs (with appropriate surface finish) at 25 degrees C and 100% RH. A comparison of strength distribution of the discs, before versus after static fatigue tests, showed no changes in strength indicating absence of slow crack growth at stresses and environment representative of manufacturing process. Such a verification is imperative for selecting appropriate surface finishes for the mirror blanks to ensure their mechanical reliability during manufacturing.

Parabolic mirror segments are produced from flat glass plates, and the glass plates are softened to parabolic surface by master mold. In the softening of the glass, the closeness of fit between the glass and the master mold will vary according to the surface tension of the glass, and it will influence the precision of the softening process. This problem brings about the necessity of improving the precision of the process.

Spin cast and fusion techniques have been available for the production of borosilicate mirrors for some time. Our efforts in pressure injection molding methodologies have been primarily devoted to developing eclectic geometric variety in substrate configuration, improved homogeneity of the medium, reduced face-plate bubbles, cross-sectional dimension control, and cost effectiveness. These attributes are intended to enhance the inherent characteristics of the existing techniques, i.e., savings in weight, high deflection, stiffness, and thermal stability. We have successfully molded mirror substrates in a diversity of sizes and shapes from 50 mm polygons through 1.4 meter rounds and 1.7 meter ellipses in a varity of aspect rations. Rib structure and mounting location is totally arbitrary and front and black plate geometry is unrestrained. In order to achieve reliable and repeatable results, it was necessary to develop new techniques in refractory design and fabrication, re-evaluate the raw media, redesign the existing furnace control and heating scenario and write new control and reporting software.

Can we make quadratic surfaces, especially paraboloidal, ellipsoidal, and hyperboloidal surfaces more simply? Are there more simple ways without using complex machines? In answer to these questions, we designed what we call the 'Cylinder Method'. It is very effective and in principle relatively simple. First, we consider the locus created by dropping a line perpendicular to the optical axis from any point on the cross section of the cylinder and rotating about optical axis. Second, we pay attention to the phenomenon that all cross sections of quadratic surfaces and cylinders are ellipses. We decide on the form of a cylinder which has a common cross section with quadratic surfaces. We discuss cylinders which are useful as tools or gauges for quadratic surfaces.

The micro-sensor field is presently proliferating with designs and approaches. We have developed a micro-spectrometer for sensing applications containing five precision surfaces, including two off-axis aspheres. The entire monolith is less than six cubic centimeters in volume. This particular design contains a bandwidth of about 2 micrometers which is centered at 980 nm. Once an appropriate starting substrate was produced, the entire system was diamond turned to maintain the required surface figure, inter-surface spacing, and surface tilts. Only three diamond turned fixtures were needed to produce the monolith. The results proved to be more than adequate for many sensing applications. Slightly altered designs could easily be produced containing different bandwidths and resolutions as needed by the customer. Due to the spectrum of interest and the fabrication method, PMMA was the material chosen for this sensor. Other designs configurations incorporating BK7 and sapphire are presently being studied.

Recent combinations of diffractive and refractive functions in the same optical component allow designers additional opportunities to make systems more compact and enhance performance. This paper describes a research program for fabricating hybrid refractive/diffractive components from diamond-turned molds using the bulk casting of sol-gel silica glass. We use the complementary dispersive nature of refractive and diffractive optics to render two-color correction in a single hybrid optical element. Since diamond turning has matured as a deterministic manufacturing technology, techniques previoulsy suitable only in the infrared are now being applied to components used at visible wavelengths. Thus, the marriage of diamond turning and sol-gel processes offers a cost-effective method for producing highly customized and specialized optical components in high quality silica glass. With the sol-gel casting method of replication, diamond-turned mold costs can be shared over many pieces. Diamond turning takes advantage of all of the available degrees of freedom in a single hybrid optical element: aspheric surface to elimiate spherical aberration, kinoform surface for control of primary chromatic aberration, and the flexibility to place the kinoform on nonplanar surfaces for maximum design flexibility. We will discuss the critical issues involved in designing the hybrid element, single point diamond-turning the mold, and fabrication in glass using the sol-gel process.

The heating process effects undesired changes of the refractive index. We must have good knowledge in this subject when appropriate refractive index should be achieved finally. Such research work was performed using two kinds of optical glass F1 and SK4. Fifteen samples of each kind had been prepared. Every one sample was heated separately to different temperature and then after the refractive index of each sample was measured. Following conclusions were drawn: for the glass F1 the value of refractive index decreases 0.0011 whereas 0.0035 for SK4--the change of refractive index occurs rather rapidly in a small range of temperature.

This paper describes a novel technique to quantitatively measure cross talk in order-packed image fiber bundles using a calibrated optical video system. At the input end of the fiber bundle, by carefully matching numerical aperture of the focusing lens, light is coupled into one single fiber. At the output end, a large area around the excited fiber is monitored using a video CCD array. The light intensity in integrated and the ratio between output light in adjacent fibers and output light from the excited fiber constitutes the percentage cross talk. Additionally, static modulation transfer function (MTF) measurements together with visual inspection was used to correlate the cross talk to the imaging properties of an optical fiber bundle. Comparison between different samples results in the following observations: typical cross talk varies between 1-15% for good versus bad imaging fiber bundles; image fiber bundles with less cross talk show higher static MTF at low spatial frequencies; different defects in the fiber bundle manifest themselves by specific signatures, both in the static MTF and the cross talk; when the cross talk is less 5%, the spatial cut-off frequency is not altered; cross talk decreases when reducing the numerical aperture of the coupling system.

A novel focus-setting technique is elaborated in terms of its theoretical analysis and experimental results. Upon the new technology a high precision instrument for measuring curvature radius of polished concave spherical surface has been developed and its practical applications are discussed. According to the appraisal made by author-itative experts, the relative measuring error of curvature radius has been reduced to the magnitude of (Delta) R/R<EQ 1/100000. The present instrument is one of key measuring devices in developing high imaging quality and large size optical instruments in advanced science and technology.

A new spherometer for measuring both the radius of curvature and focal length based on the self-reference shearing interferometric technique is described. Experimental results and error analysis are presented, and both of them show that a high degree of accuracy can be achieved using this spherometer.

Optical angle gauge is one of the measuring instruments standardized by the National Metrological System of Chian for verification of plane angle. In this paper a new method for verifying the optical angle gauge is described and a related photoelectric goniometer developed by ourself is introduced. As a datum instrument for measuring deviation angle of optical wedge, its maximum error is less than 0.1 inch. In addition, some examples of its application for calibrating indication error of various angle measuring devices are presented, and its distinctive advantages, especially in calibrating 2D autocollimators, are demonstrated.

The primary mirrors for the ESO very large telescope project consist of four f/1.8 8.2-m diameter actively supported Zerodur menisci, 175 mm thick. Their optical figuring is under way at REOSC (France) plant. Two mirror blanks have been delivered to the optical manufacturer and the two remaining at the blanks' manufacturer (SCHOOT Mainz, Germany) are due for delivery by the second hald of 1995. Th polishing of the first Zerodur mirror blank is due for completion in the third quarter of 1995. The REOSC optical fabrication and testing methodology will be presented, and the current results will be compared to the specified performance.