Basic interferometry and holographic interferometry are becoming useful tools for precision measurements. The use of electronic solid state detector arrays, together with, small computers for extracting the information from the interferograms improve their applications. Computer analysis is becoming increasingly important. A lot of information can extracted from the interferograms leading to higher sensitivities and accuracies.1 New technologies to generate optical surfaces have been introduced. Progress has been made in manufacturing of metallic spherical and aspherical surfaces using diamond turning techniques, for instance. The improved production techniques will motivate the optical designer to use aspheric surfaces more frequently especially for IR applications. Using replica techniques, aspheric surfaces will be used more often in serial production of components used for the visible or near IR. In consumer products the cost of precision optics is relatively high. Using inexpensive aspheric lenses could cut the cost considerably. The application of aspherical optical surfaces could be helpfu) to reduce the number of optical elements reducing the total weight of the optical system and possibly the cost, in addition, the quality may be improved. Replication of an aspheric surface on a thin film onto a spherical glass body using photocurable la cquer is very cost effective.2 The aspheric mold as well as the resulting aspheric optical elements need to be tested. Optical contactless testing techniques using computer generated holograms can be very useful for industrial applications for micro- and macro structure analysis of aspheric surfaces.3

The advent of automatic data processing for interferometry greatly reduced the complexity of interferometric testing, thereby significantly increasing its usage both within and outside of the optics community. A great many of these new applications require special data processing and output data not normally used for testing lenses. To address these new measurement problems many data analysis programs were written ranging from advanced analysis of wavefronts (i.e., Zernike Polynominals, Point Spread Function, Modulation Transfer Function) to analysis of mechanical surfaces such as Winchester disk read/write heads. Some of these programs were very specific to a particular application and some tried to be general and as such became cumbersome. Even with much of this software available for sale to the general public most applications other than simple surface and lens measurements usually cannot be solved directly using the available programs. In an attempt to solve this problem, we have developed software for interferometry that allows users to easily develop their own measurement routines. The solution was to take a version of the BASIC programming language and add the commands necessary to do interferometry. This software is resident in a processor that can easily be adapted to a large number of interferometry applications. By using this processor and its associated software with an appropriate interferometer, it is possible for the user to tailor the measurement to a particular application. This can be very useful in an optical production shop where each different testing application can have its own program. If the program is written properly, the operator will not have to set up any default conditions or format the output; the operation of the program can be reduced to the pushing of a single button and the output will be formatted properly for that particular test. Examples of this system in actual optical shop testing situations will be discussed.

Phase-measurement algorithms for calculating the phase of a wavefront from interference fringe data are compared. Experimental data show that different algorithms yield different phase values when using the same intensity data. A computer simulation of errors due to phase-shifter miscalibration and nonlinearity, as well as detector nonlinearity is performed to show that certain algorithms are more sensitive to some errors than others. Dependences of each of these errors is found versus percent of error over a 2ic range of phase values. These results enable the determination of what system errors are present in a phase-measurement interferometer.

A variety of parameters are currently used to describe the performance of phase-measuring interferometers. Unfortunately, there is very little documentation concerning their definitions, quantifications, and suitabilities. This paper endeavors to address these issues as well as to discuss the theoretical foundations and limitations of specific parameters in hopes of stimulating progress towards a definitive set of phase-measuring interferometric performance parameters.

A compact, no-moving-parts, dc-phase-measuring interferometer has been designed and manufactured. This interferometer was designed to provide high-density sampling, high-accuracy wavefront measurements under adverse conditions. A crossed grating shearing interferometer technique has been utilized where all data is simultaneously collected on a single fixed detector array. This highly stable configuration provides a means of obtaining phase information using cw, shuttered (1 to 10-3 sec), or ultra-short-pulsed sources. This self-contained instrument can be used to take large numbers of interferograms quickly and reduce their data efficiently to obtain wavefront constructions.

The collimation tester is one of the most simple devices available for examining optical wavefronts. Based on a shearing interferometer, its sensitivity can be adjusted as required. Methods of use in collimating laser beams will be described, in addition to other applications involving planarity. Versatility in wavelength coverage will be discussed.

Two important problems in interferometric aspheric surface testing are introduced and approaches for overcoming the problems are discussed. Firstly, the definition of the system aberrations of an optical system is given. The effect of system aberrations is explained by the result of computer simulations. An approach for the reduction of their influences using third order aberration representation of them is discussed. Secondly, the problem in the interpretation of an interferogram is discussed, and an interpretation method based on Huygens' principle is discussed. By the use of these methods, the influence of system aberration can be reduced, and the exact shape of aspheric surface to be tested can be obtained from the phase distributions measured.

Itek Corporation has been awarded a contract for the manufacture of the primary mirror segments for the W. M. heck Observatory. The primary mirror will contain 36 hexagonally-shaped off-axis aspheres which, when aligned, will form the collecting area of a 10-m-diameter mirror. To ensure integration to the telescope, each segment must be the correct part of the desired parent asphere. During fabrication, careful optical and dimensional metrology must be performed on each segment. This paper outlines the in-process and final metrology plans for each segment.

An optical testing facility has been established for testing large-aperture long-focal-length diamond-turned optical components. This paper discusses some of the problems and limitations of testing such optics as well as techniques to minimizing their effects. Illustrative results will be presented.

A theory describing the intensity distribution in out-of-focus images of a telescope mirror is derived based on geometrical ray optics for a rotationally symmetric optical system. Surfaces at which the intensity distribution diverges are shown to be the same as the classic caustic surfaces. The contrast variations at out-of-focus positions are proportional to the square of the F-number of the system. For a large F-number system such as a telescope, the out-of-focus intensity distributions provide a sensitive method for determining zonal error distributions of the optical system.

An interfermetric systm capable of precisely and automatically measuring an aspheric surface is described. Major techniques are lateral-shearing interferometry using the fringe scanning detection method and automated alignment. Key elements of this system are a 1ateral shearing interferometer with a beam-splitter and a piezoelectrically-driven corner-cubes-prism, a precise alignment mechanism driven by an output signal of a sensor which detects an object location, and a computer for analyzing fringe information and controlling the system. Rms accuracy of λ/32 is achieved from the evaluation on a convex parabolic mirror with asphericity of 200λ and a curvature radius of 35 mm.

Interferometric techniques using null correctors and computer-generated holograms have become the conventional methods of aspheric surface measurement. These techniques involve substantial design and manufacturing time prior to testing. Additional complexity is added by the necessity to measure all optical surfaces, air spaces, and refractive index of the components. This is obvious for a null corrector, but even when using a computer-generated hologram an interferometric objective is used to reduce the curvature of the wavefront returning from the test surface. The air spaces, indices and surfaces of the objective must also be known as accurately as those of a null corrector. The original computer-generated artwork for the hologram must also be accurately produced and reduced using optics with very little distortion. Both of these techniques, therefore, become prone to systematic error due to inaccurate measurement of the components of the test system. Mechanical profilometry is an alternative to these techniques. The accuracy of the system can be verified using known spheres because the spheres will be measured by the same process as the asphere. In the process of measuring the asphere, a meridian of an optical surface can he rotated on its best fit sphere or measured relative to a straight line. A mechanical gauge is used to measure the deviation from the line or best fit sphere. Accuracy of one-tenth wave is possible and accuracy of one-quarter to one-half wave is routine. The mechanical profilometers are quite universal and require very little equipment specific to a given test surface. This reduces the front-end time for specific measurements compared to null correctors and holographic nulls. This paper discusses the geometry of aspheric profile measurements and instrumentation. It includes a description of some transducer options, mechanical considerations, calibration, sample surface figure error curves, and a brief discussion of potential errors.

Although a worrying problem ever since optical components were first produced, the measurement of imperfections on optical components has become more acute in the last few years due to an increase in international trade, the more widespread use of laser optics and the trend towards automation in component production. Quite apart from their functional effect on certain classes of component, visible signs of surface damage have a direct impact on perceived quality and therefore customer acceptability of the product. This paper reviews previous attempts to quantify scratches and digs in relation to a variety of different so-called standard artefacts and discusses the extent to which these attempts have met with the approval of quality engineers and customers and are therefore likely to form a satisfactory basis for embodiment in an international standard. The results of recent surveys, supported by laboratory trials, conclude that all current national standard procedures dealing with surface imperfections produce an uncertainty of measurement in excess of levels normally regarded as acceptable by quality engineers. The reasons for this large uncertainty of measurement are discussed and a suggestion made for a new approach to the measurement of scratches based on photometric rather than geometric considerations. The reactions of potential users and the results of measurements of uncertainty using this approach indicate that it has much to offer both as a means of training inspectors and as an inspection tool for workshop use.

The economical (fast) fabrication of large aspherics that have departures from the nearest sphere ranging up to several millimeters requires aspherization by grinding methods rather than polishing methods. One grinding method is the high-precision machining of the aspherics with fixed diamonds on a uniaxially controlled machine. Following aspherization, polishing can be accomplished with suitably flexible tools, ideally in conjunction with a computer-controlled polisher. The Optical Sciences Center has in operation two such machines: the Large Optical Generator (LOG) and the Swing-Arm Generator (SAG) . Both machines, relying on quite different tool-path geometries, can cut large and fast aspherics to about 1-ttm RMS surface accuracy, which brings the surface to well within the capability of polishing tools to reach typical final-figure accuracies. To take full advantage of this machine accuracy, the resulting surfaces must exhibit sufficient smoothness to minimize the post-generation fabrication effort. Clearly, the less surface damage the less effort will be necessary. However, the surface finish is not the most important consideration when determining the next step. Rather, it is the depth of the subsurface damage caused by the diamond. This damage must be removed in subsequent operations to produce a surface having the least amount of scatter and the highest mechanical strength achievable. During the fabrication of a 40-in.-diameter Zerodur off-axis parabola on the LOG and a 27-in. Zerodur off-axis parabola on the SAG, we evaluated the subsurface damage throughout the generating process using the contact method described below for each grade of diamond that was utilized.

Surface finish has vertical and transverse attributes. In optics the first is characterized by the root-mean-square roughness, and it is natural to seek an analogous length parameter, or "correlation length", to characterize its transverse properties. This paper evaluates various candidates for this purpose by examining their robustness to the finite-bandwidth effects inherent in all real measurement processes. The critical consideration is the magnitude of the correlation length with respect to the range of surface wavelengths in-cluded in the measurement: When the length parameters of the surface fall outside the measurement bandwidth it is possible to get totally spurious values for the measured parameters, and when they fall inside the band-width, the measured and true values can still differ by significant factors. This sensitivity of the correlation function to bandwidth effects arises from the fact that the measurement process modifies it by convolution; in contrast to the finish power spectrum, where the modifications involve simple multiplication. The present results argue against the use of length parameters derived from the measured correlation function as precise measures of the transverse character of surface finish, although they may still be useful for compar-ative and diagnostic purposes. The most precise characterization of the combined vertical and transverse properties of surface finish are given in terms of its power spectral density.

A method for extending the measurement range of a two-dimensional optical profiler is described. Using a phase-modulated Mirau interferometer coupled to a microscope with a computer controlled stage a high density of data points is obtained over an extended region. Successive interferograms are combined by matching the piston and tilt of each scan to the proceeding scan. This instrument provides surface roughness measurements for a wide variety of surfaces which are quick and non-contact. This paper examines important features of the profiler as well as presenting data of the rms surface roughness of several substrates.

The objective of this work was the development of a noncontact method for the measurement of the wavyness of polished surfaces. To obtain simplified measuring procedures interferometric methods were abandoned in favor of a focus seeking system that was developed earlier for machined surface inspection. The system is based on differential detection of the reflected light through two properly prepared and adjusted filters. The filter structure and alignment were investigated to obtain linear response and maximum sensitivity that amounted to 20nm in preliminary experiments.

Optical measurements of surface topography require the translation of measured phase shifts, Φ , into distances, Z , using a physical model for the air-surface interface. The usual expression used for this purpose = etp is based on the simple-interface model: air or vacuum above and a homogeneous half space of material below, with a sharp dividing line in between. This paper discusses the corrections to this expression which appear when one considers more complicated models. These appear as added pseudoheight contributions which arise from spatial variations of the parameters appearing in the more camplicated models, in addition to the geametrical height variations included in the simple-interface model. Formulas are given and results illustrated for several cases of practical interest.

A WYKO TOPO-2D optical profiler system with four different magnification objectives (2.5X, 10X, 20X and 40X) was used to measure the microroughness properties of a set of five different flat samples. Measurements are compared in terms of the average periodogram for each surface as a function of microscope objective, and the effects of correcting for the transfer function of each objective are assessed.

This paper compares the WYKO and Sommargren noncontact surface profilers, describes their utility and limitations, their accuracy and their use in polishing research. Surface readings below lA rms by these instruments are reported.

The scattering of light from a rough surface of a semi-infinite medium is analyzed by means of a Green's function formalism based on the macroscopic Maxwell equations. It is shown that every direction of scattering in the half-space above the surface, in the small roughness limit, is connected to precisely one spatial Fourier-component of the topography function. The possibilities for reconstructing the topography by ellipsometric measurements of the vectorial amplitude of the electric field scattered diffusely outside the specular direction are investigated. Using laser light having wavelengths of 458 nm or 514 nm, scattering measurements are performed on a polycrystalline AQ surface. Results are presented for experimental setups with two different angular resolutions.

The quality of the polish of an optical surface can be evaluated by the amount of light it scatters. Some optical materials, notably Zerodur' and CervitTM, exhibit bulk scatter; that is, light can be scattered from the interior of the material as well as from the surface. The presence of bulk scatter makes it difficult to observe the surface scatter, especially when the level of surface scatter is very low as is the case with well-polished optical surfaces. A novel instrument has been built that can distinguish between surface and bulk scatter by differences in their polarization characteristics. The instrument will give zero output signals for bulk scatter while being sensitive to the surface scatter. The instrument can be said to be blind to bulk scatter.

The use of two-dimensional Fourier optical signal processing has been applied to the field of surface defect detection. Both transmissive and reflective objects can be analyzed using this system. The object is placed in a coherent optical processing system. By selectively blocking information in the frequency spectrum which contribute to background light and noise levels, only information relating to physical defects in the test object will be observed at the output plane of the system. The system has been observed to be relatively impervious to mechanical vibration, allows for the inspection of large areas instantaneously, and has relatively good resolution.