A simple method for interpreting Hartmann test data during the figuring process of a concave asphere is presented. The procedure is relatively fast and permits an estimation of the degree of smoothness and rotational symmetry of the mirror. It is not necessary to use all of the data point on the screen, but only those near one of it's radius. After the results from this test are satisfactory, the traditional, more complete and time consuming methods may be employed.
Three null correctors of the Offner type were designed to test an F/1 paraboloid and their performance compared to determine the optimum configuration with reference to wave-front residual, tolerancing, alignment and fabrication difficulty.
A new phase-shift method based on the spatial phase modulated carrier fringe pattern is presented in this paper. The four phase-shifted fringe patterns, in which the phase differences between the adjacent pattern are n12, can be generated from only one fringe pattern, and all of the procedures can be done automatically. This method can be used in the condition of incoherent metrology and dynamic measurement.
Very accurate and detailed testing of small and moderate size optical surfaces can be realized by phase shifting interferometry. However, for the very large primaries needed for the next generation of telescopes, it is difficult and expensive to keep vibrations low enough for this method to work. A modification to the usual phase shifting interferometry reduction algorithm permits measurements to be taken fast enough to essentially freeze out vibrations. Only two interferograms are needed with an exact phase relationship; and these can be recorded very rapidly on either side of the interline transfer of a standard CCD video camera, prior to charge transfer readout. The third required interferogram is a null. In the developed implementation, two frequencies, dv/v.10-8, are generated with orthogonal polarizations. A Pockels cell rapidly switches the frequency entering the interferometer, resulting in a phase shift over the long path difference of the interferometer. The two time critical interferograms are acquired with a lms separation resulting in a reduction in sensitivity to vibration of one to two orders of magnitude. Laboratory tests were performed to compare this "2 + 1" system with a commercial phase shifting package. Similar phase determination accuracies were found when vibrations were low. However, the "2 + 1" system also succeeded when vibrations were large enough to wash out video rate fringes.
Frequency spectra are analyzed for two typical speckle-shearing interference recordings - single-aperture speckle-shearing camera (SASSC) and double-aperture speckle-shearing camera (DASSC) to show the fringe variation due to the displacement involved in differentiation. A speckle statistics is then made for filtering system to show the fringe contrastvariation resulting from the object deformation and the aperture of filtering-image system.
The effects of the measurement transfer function in a Wyko-like profiling microscope have been explored analytically, and large attenuations found over the upper half Hof the optical bandpass. In the case of corrugated surfaces these effects can be eliminated using a universal inverse-filtering routine. In the case of isotropically-rough surfaces the effects are larger, and although restoration is possible, no universal filter exists.
A functional integral representation of the statistics of rough surfaces is developed. The assumption of locality, defined in the text, produces a general form for the probability functional which automatically contains the mean square height and correlation length of the surface. The correlation function to lowest order is predicted to be a K0 modified Bessel function for all rough surfaces. The power spectrum obtained from this modified Bessel function Is In good agreement with the measured power spectra of rough surfaces. Fractal behavior occurs naturally due to the anomalous scaling of the correlation functions when higher order terms are included in the calculation.
An optical-digital approach of roughness classification is described that uses the irradiance moments extracted from the Hartley transform intensity of an input object. A set of discriminant features based on the lower moments are done for typical machined samples. An analysis of these parameters in surface classification is also carried out.
One of the principal concepts of sub-Nyquist interferometry is the use of all available knowledge about a surface to extend the available measurement range. This paper will explore the use of this a priori information in resolving the ambiguities found in interferometric step-height measurements. Information about the height of a step that has been obtained by another measurement technique or through process parameters, such as etch rate, may be used for this purpose. The theory behind this type of measurement and experimental results are presented.
Step heights measured with an interference microscope in monochromatic light are generally limited to steps not exceeding 1/4 of the illuminating wavelength due to the difficulty in following the fringe order across the step. There are several ways to overcome this ambiguity. One method tracks the fringe contrast across a step to help determine the absolute phase across the step height.1 Another method uses two measurements at differing wavelengths (i.e., two-wavelength interferometry) to create a synthetic wavelength larger than either single wavelength.2 The quarter-wave limit of the synthetic wavelength then being significantly larger than either single wavelength measurement. The method I propose utilizes the fact that the fringe spacing in interference microscopes with extended source illumination is greater than half a wavelength3 (i.e., 1 fringe width = λ/2•obliquity factor). The obliquity factor varies as a function of the illumination numerical aperture and the distribution of the intensity of the illumination across the pupil of the microscope objective. Changing the aperture stop setting in the illumination arm of the interference microscope between successive measurements changes the magnitude of the obliquity factor, thereby creating two single effective wavelengths. Two-wavelength interferometry can be accomplished by the creation of a larger equivalent wavelength from the two single effective wavelengths; that is, from one wavelength multiplied by two different obliquity factors.
The numerical aperture (NA) of a microscope objective can affect the measurement of surface height profiles. Large NA objectives measure height values smaller than the actual values. An experiment to calibrate these effects on objectives with NAs of 0.1 to 0.95 is described using four traceable step height standards and a computer-controlled interferometric optical profiler utilizing phase-measurement interferometry techniques. The measured NA scaling factors have good agreement with a theory developed by Ingelstam and indicate that the effective NA rather than the nominal NA is the important quantity. NA scaling factors are determined to an uncertainty of ±1% for NAs of 0.5 or less and ±2% for NAs of 0.9 or greater.
Several systems have recently been developed that apply various modulations onto the optical beams in order to both enhance the contrast from particular features and to reduce the effects of noise and microphonics. Consequently different imaging responses may be obtained at different photodetector output frequencies. This paper analyses a selection of AC profilometer systems by extending the analysis used for the conventional scanning optical microscope.
We have imaged diamond turned gold surfaces and a gold coated silicon surface with the STM. In order to determine the reproducibility of the topographic information obtained with the STM, we imaged the same diamond turned gold surface with different tips. Both mechanically formed and electrochemically etched tips were used. Surface images observed with the STM varied from tip to tip for both methods of preparation. The use of the STM to image optical surfaces hinges on the ability to manufacture stable and reproducible tips.
In the past two years scanning tunneling microscopy (STM) has expanded rapidly in several areas, over the size range from looking at atoms to looking at integrated circuits. In this paper we will cover some of the recent uses of STM and show some of the three-dimensional surface images which have now become routine.
This paper describes the implementation of an automatic focus control system (Autofocus) to an optical non-contact surface profiling instrument. Such a focus control system extends the operation of the surface profiler to measurements of both roughness and figure that would otherwise be limited by the optical systems depth of focus. Measurement results are presented from a variety of curved samples.
An interferometric optical profiler using a Videk Megaplus camera interfaced to an 80386-based computer has been developed for measuring surface roughness and three-dimensional profiles with high spatial resolution. The camera has 1320 x 1035 square pixels which are 6.8 gm on a side. Fields of view over 1024 x 1024 pixels can range from 35 gm to 4642 gm with spatial sampling intervals of 0.034 pm to 4.5 gm. The spatial resolution is limited by the optical resolution of the system. The computer system consists of a 80386 microprocessor and a floating point coprocessor, an AT compatible IQ bus, and a VGA monitor with software running under Xenix 386. Digital data from the camera is captured using a frame grabber and displayed as live video on a monitor at 7 frames per second. The rms repeatability of measurements made with this instrument is better than 0.7 nm and can be half this value with averaging. This paper describes the system and its performance along with results of measuring various samples such as gratings, diamond turned optics, magnetic media, recording heads, and plastics.
The advantage of optical profilometers based on a common path type interferometer is their insensitivity to microphonics. This means that such systems are capable of very accurate phase measurements. In this paper we will present a system with this characteristic. Two beams are directed onto the sample along the same path. One beam is focused onto the sample surface and the second remains collimated, acting as a large area reference. This is achieved using a specially constructed lens which allows the relative size of the illuminated areas on the sample to be chosen arbitrarily, giving a large reference and a high signal to noise ratio. In most existing systems this ratio is not independent and there is a trade-off between the effective ratio of the spots and interference efficiency. In our system no such trade-off exists. Each reflected beam is interfered with a third frequency shifted beam. They are then detected separately, resulting in two AC signals. The phase of these signals are then compared to provide the object surface phase structure. Path length fluctuations due to microphonics are cancelled by this comparison. Results obtained from a prototype version of the system will be presented.
Since 1912 X-ray diffraction has been the principal method for determining the crystal axis orientation of materials. This test is generally time consuming and must be done in a shielded location. A new real time method has been developed using a rapidly scanned laser beam to analyze the surface morphology of the crystal. By relating the minute facets in the surface to the underlying lattice structure, the crystal axis orientation can be quantitatively determined. The laser beam rapidly rotates about the axis of the instrument projecting a narrow beam at a small point on the surface of the crystal. This scan illuminates the crystal from all azimuth angles while the angle of incidence is varied in successive incre-ments. The laser light reflected from the facets is detected by a centrally located sensor and correlated with a precise clock signal to determine its angular position. The accumulated data is then processed by appropriate algorithms to determine the crystal axis orientation. Then with the aid of individual solid geometry algo-rithms, the deviation of the crystal axis is calculated. This is accomplished in one second and the results are displayed on the monitor in degrees and tenths of a degree. The laser scanner is fully controlled by a computer which is also used to reduce the raw data. The data display is presented on the monitor screen in sufficient detail to determine subtle differences in crystal structure and orientation. Measurements have been made on face centered cubic single crystals of semiconductor materials such as gallium arsenide, cadmium telluride and silicon in 1,0,0; 1,1,0; and 1,1,1 orientations.
In the present paper, cavity shapes are profiled using two opposing collinear panoramic annular lenses. A speckle pattern is projected out from one panoramic annular lens onto a reference stan d or onto lle cavity wall itself, and the resulting speckle distribution is recorded through e second panoramic annular lens. Since the speckle pattern moves as the shape of the cavity changes, H e variation in cavity shape can be measured by applying correlation techniques to the speckle patterns obtained before and after the cavity shape is modified. The apparent speckle movement is computed by numerically correlating small subsets extracted from each pattern. These shifts are used to measure surface deflections or to contour the cavity with respect to a reference shape.
The existing British standard, BS 4301:1982, relating to cosmetic defects, such as scratches and digs, has proved difficult to apply. Both industry and the Ministry of Defence have stressed the urgent need for a technique which can be readily employed by the majority of optical component manufacturers, is suitable for assessing appearance and functional flaws, and has traceability to national standards. This paper describes a new proposal, under consideration by a BSI committee, which recommends replacing standard scratch artefacts, made by etching a rectangular trough in a glass substrate, by the use of a standard line reference plate used to calibrate a flaw comparator instrument. The problems currently met by industry when using present visibility standards, such as BS 4301:1982 and MIL-0-13830A:1963, will be discussed in relation to the new proposal in which the signifance of a flaw is quantified in terms of a line-equivalent width (LEW) rating. The method of operation of a microscope image comparator, used for calibrating reference flaws in terms of their LEW value, will be described, and results recently obtained by inspecting transmitting and reflecting surfaces, which may be flat or curved, coated or uncoated, will be presented.
In this paper, I present a history of the scratch-and-dig standard for optical surface quality and show that this standard has since its inception been recognized as a cosmetic standard and not as an objective or performance standard. In addition, I attempt to dispel the myth that the scratch standard was changed during the 1960's and show that scratch number cannot be related to scratch width. Finally, I describe a preliminary aging experiment that suggests that the scratch standards have not aged with time and are, in fact, extremely stable.
confocal Laser Scanning Microscope has been developed which has many advantages over conventional microscopes in many different applications including small spot profilometry, CD measuring and 3D surface profiling.
A recently-developed long-trace surface profiling instrument (LTP) is now in operation in our laboratory measuring surface profiles on grazing incidence aspheres and also conventional optical surfaces. The LTP characterizes surface height profiles in a non-contact manner over spatial periods ranging from 1 meter (the maximum scan length) to 2 mm (the Nyquist period for 1 mm sampling period) and complements the range of our WYKO NCP-1000 2.5X surface roughness profiler (5 mm to 9.8 pm). Using these two instruments, we can fully characterize both figure and finish of an optical surface in the same way that we normally characterize surface finish, e.g., by means of the power spectral density function in the spatial frequency domain. A great deal of information about the distribution of figure errors over various spatial frequency ranges is available from this data, which is useful for process control and predicting performance at the desired wavelength and incidence angle. In addition, the LTP is able to measure the absolute radius of curvature on long-radius optics with high precision and accuracy. Angular errors in the optical head are measured in real time by an electronic autocollimator as the head traverses the linear air bearing slide. Measurements of kilometer radius optics can be made very quickly and the data analyzed in a format that is very easy to understand.
Where surface quality is crucial, metal manufacturers employ surface inspection systems that use image processing techniques to distinguish defects from normal background. These systems are well suited for surfaces in which the defects are characterized by sharp variations in intensity from the average background intensity. However, in the presence of tool marks, the ability of the surface inspection systems to distinguish defects from normal background is severely degraded. This paper presents techniques based on optical/image processing concepts to suppress the tool-marks and thereby enhance the detectability of surface defects.
A surface analysis system commonly used in semiconductor processing is used to characterize microcontamination generated by optical thin film deposition processes. The system is first described and an overview presented of its operation and application when analyzing bare silicon wafers. Bare silicon wafers were used as witness plates to sample microcontamination generated by operations in ion beam sputtering and electron beam evaporation thin film processes. Additional wafers were used to test cleaning methods. The use of the surface analysis system was found to be invaluable in rapidly reducing point defects generated during processes, assessing the cleanliness of work environments, and providing feedback for developing cleaning processes.
To verify the Starlab Mission's optical requirements for a pressurized module window, a program with a series of specialized optical tests was required. In this paper a brief description of each optical test will be defined along with the description of the Spacelab Optical Viewport (SOV) glass assembly. The glass assembly is a window installed in the aft end cone of the Spacelab module. It has been designed to act as a structural member in the module shell and also to provide a path through which radiation can pass. The optical test definition will include the test requirements, test description, and test results. A transmitted wavefront and transmission test was conducted to verify the wavefront error and glass coating quality respectively. A Bidirectional Reflectance-Distribution Function (BRDF) measurement was taken to determine the outgassing effects of the window's non-metalic components and cleanliness level of the glass from assembly to installation in module. The optical performance of the assembled unit was verified by performing a wavefront and dispersion test at ambient conditions and wavefront test under simulated space flight pressure. This test program satisfies Marshall Space Flight Center (MSFC) requirements for optical space flight hardware. The SOV glass assembly design has gone through preliminary design review, critical design review, and has been fabricated, assembled and tested.
The performance of printers and copiers often depends in part on the surface topography and roughness of cylindrical components used in paper handling, cleaning, toner deposition, and electrostatic image formation. This paper describes the development and evaluation of a reliable, inexpensive optical technique for rapidly estimating the roughness of an ablated, thick polymer film on an elastic, cylindrical substrate. The intended application is to component inspection and process control in a manufacturing environment.
Most optical systems currently in use utilize a single type of measurement from which deductions are made about sample properties such as surface profile or reflectivity variations. In this paper we will present a system which can perform several independent measurements simultaneously. A Bragg cell is used to divide an incident light beam into two probing beams which are focussed onto the sample as two close spots. Independent measurement of differential phase and intensity is enabled by virtue of the Bragg cell introducing a frequency shift between the two beams and modulating the two in phase quadrature. By varying the frequencies in the Bragg cell drive, the separation of the spots may be changed, giving the system two distinct modes of operation, for imaging and metrology. The system is common path thereby giving considerable improvement in sensitivity over conventional interferometers which are limited by mechanical vibrations. A thorough description is given of the system which, together with computer simulations of the acoustooptic interaction, is used to quantify the performance of the system and to outline design considerations. Throughout the paper, emphasis is placed on identifying and describing problems which may be encountered in a practical implementation of the system.
A new imaging process of laser speckle has been proposed especially for characterizing and measuring the change of the surface roughness of metallic samples, as a result of corrosion attacks in aqueous solution. The process could provide useful microscopic information about a preceding corrosion attack, i.e., pitting corrosion, that may take place at the metal surface. In other words, one may predict from this process different kinds of corrosion forms before actually can be detected by the naked eyes. In addition, a correlation between data of surface roughness and corrosion of metallic surfaces can be established based on the proposed process. In the present report, preliminary results on a stainless steel examined in a saline solution are given.