Optical scatter is a bothersome source of optical noise, limits resolution and reduces system throughput. However, it is also an extremely sensitive metrology tool. It is employed in a wide variety of applications in the optics industry (where direct scatter measurement is of concern) and is becoming a popular indirect measurement in other industries where its measurement in some form is an indicator of another component property - like roughness, contamination or position. This paper presents a brief review of the current state of this technology as it emerges from university and government laboratories into more general industry use. The bidirectional scatter distribution function (or BSDF) has become the common format for expressing scatter data and is now used almost universally. Measurements made at dozens of laboratories around the country cover the spectrum from the uv to the mid- IR. Data analysis of optical component scatter has progressed to the point where a variety of analysis tools are becoming available for discriminating between the various sources of scatter. Work has progressed on the analysis of rough surface scatter and the application of these techniques to some challenging problems outside the optical industry. Scatter metrology is acquiring standards and formal test procedures. The available scatter data base is rapidly expanding as the number and sophistication of measurement facilities increases. Scatter from contaminants is continuing to be a major area of work as scatterometers appear in vacuum chambers at various laboratories across the country. Another area of research driven by space applications is understanding the non-topographic sources of mid-IR scatter that are associated with Beryllium and other materials. The current flurry of work in this growing area of metrology can be expected to continue for several more years and to further expand to applications in other industries.

We discuss scattering in the context of the Stokes vectors and Mueller matrices that characterize the interaction. In order to study surface structures using light-scattering techniques it is useful to examine the nature of light scattered from perfect and perturbed mirror surfaces.

This study of light scattered by sinusoidal surfaces shows that such a configuration can be used as a material standard to help calibrate instruments that measure the BRDF of arbitrary surfaces. Measured and computed values of the power scattered into the diffraction peaks show good agreement, and such calculations can be further improved and used to verify the standards.

The bidirectional transmittance distribution function (BTDF) of two sets of scratch/dig standard sets were measured. These sets were representative of the inspection standards used in the optical industry to characterize polished surface defects. Measurements were taken with a small (1 mm diameter) illumination beam to maximize signal. The increase in average BTDF that results from a single scratch or dig over the MIL-STD 20 mm diameter surface was then calculated to determine what overall impact a defect will have on system stray light above base surface scattering due to surface micro-roughness. A BTDF measurement was taken with the illumination beam centered on the defect, then with it centered on a smooth section of the reference sample to find the increase in scattering caused by the defect. Results show that dig scattering, when normalized to account for the single dig per 20 mm MIL-STD inspection area criteria, did not catastrophically increase the 0.05 B0 (B0 is the BTDF at 0.57 degree(s)) at 633 nm characteristic of a high quality optical surface. As intuitively expected, dig scattering was angularly symmetric. Scratches, however, scattered highly directionally. Normalized BTDF is substantially increased from a smooth surface's typical 0.05 B0 perpendicular to the scratch axis, but is unaffected in other angles. On average, the scratches may not have increased net surface scattering. Scattering from the defects on the surfaces below the 40-20 scratch/dig level was found to not cause a catastrophic increase in scattering over the level as a well-polished optic (typically 4 angstroms rms roughness). Since comparisons with scratch/dig samples only serve to provide a measure of the localized defects, and fail to be useful in determining the low-level scattering from the surface microroughness, one should not assume that a '40-20' surface is necessarily a low-scattering optic.

It is finally being recognized that residual surface roughness over the entire range of relevant spattalfrequencies must be specified and controlled In many precision optical systems. This includes the 'mid" spatial frequency surface errors that span the gap between the traditional "figure" and "finish" errors. This is particularly true for enhanced reflectance multilayers if both high reflectance and high spatial resolution are desired. If we assume that the interfaces making up a multilayer coating are uncorrelated at high spatial frequencies (microroughness) and perfectly correlated at low and mid spatial frequencies, then the multilayer can be thought of as a surface power spectral density (PSD) filter function. Multilayer coatings thus behave as a low-pass spatial frequency filter acting upon the substrate PSD, with the exact location and shape of this cut-off being material and process dependent. This concept allows us to apply conventional linear systems techniques to the evaluation of image quality, and to the derivation of optical fabrication tolerances, for applications utilizing multilayer coatings.

It has been demonstrated that erosion of sensor window surfaces can be measured on the flight line with reflective scatter instrumentation. At issue is the relationship between these damage sensitive measurements and the corresponding loss of system performance. Rain and sand erosion of IR sensor windows can limit system performance in three ways. In the case of IRST's, background scatter from window defects increases the system noise floor which limits range. Image resolution degrades in FLIR instrumentation as window erosion increases. Finally for both systems, severe damage can cause window breakage resulting in loss of the sensor system and possibly the aircraft. This paper reports the results of initial studies that correlate reflective scatter measurements to the loss of mid-IR performance. High angle and near angle transmissive scatter from window damage are responsible for different types of system degradation. Both are studied and related to reflective scatter measurements.

Light which is emitted from a point source and is randomly scattered by an ensemble of particles is found to form an image of the point source. Results are presented for light which is emitted from a 25 micrometer pin hole and is then scattered by a suspension of .296 micrometer polystyrene spheres. This phenomenon is interpreted in terms of coherent multiple backscattering. The present study investigates the effects of particle density on the image resolution.

One of the most interesting phenomena associated with the scattering of light from a randomly rough surface is that of enhanced backscattering. This is the presence of a well-defined peak in the retroreflection direction in the angular distribution of the incoherent component of mean scattered intensity of the light scattered from such a surface which is due primarily to the coherent interference of each multiple reflected optical path with its time-reversed partner. It is an example of a broader class of multiple scattering phenomena that goes under the name of weak localization. Not all of the manifestations of weak localization in the interaction of light with a randomly rough surface are in backscattering. It was recently shown that the average diffuse intensity from randomly rough surfaces with even symmetry can be enhanced or reduced in the specular direction due to the constructive interference between correlated pairs of scatters. In this paper we will present recent theoretical analysis and experimental results that cover four kinds of enhancement: the enhanced backscattering, the enhanced transmission, the enhanced specular, and the enhanced refraction for 1-D and 2-D surfaces. These are manifestations of coherent effects that remain after ensemble averaging.

This paper discusses basic issues involved in the estimation of surface spectra from laboratory measurements, the development of physically-based spectral models, and the estimation of finish parameters and their associated errors.

An electroless nickel plated over aluminum mirror was tested for BRDF and surface profiles at two stages: First, after a standard polish for optical figure, and second, after a `super finishing process' which is designed to minimize optical scatter. BRDF measurements corresponding to spatial frequency range between .0016 micrometers -1 to 1 micrometers -1 were obtained using 0.633 micrometers , 1.06 micrometers , 3.39 micrometers and 10.6 micrometers lasers. The conversion formula used to derive PSD2D from BRDF data is based on the Rayleigh-Rice vector theory. Measurements of the same spots by an optical profilometer (WYKO TOPO-2D) with several objectives were used to cover similar spatial frequency limits. The surface finish statistics, extracted from the profile data, was processed to produce PSD1D. Then, composite PSD1d were fitted to an analytic function for PSD1D using the K- correlation model approach. The derived PSD2D from profile data was readily determined from the A,B,C coefficients associated with the K-correlation model. The derived PSD2D from BRDF and profile measurements were compared to quantify the difference in surface finish statistics between the standard polished mirror and the same mirror after `super finishing.' Good correlation is found for the standard polished surface. This corroborates the `topographic scatter' behavior for the clean, standard polished nickel surface which promise useful interpolation of BRDF values to wavelengths at which direct BRDF data is unavailable. Correlation for the `super finished' surface is worse, probably due to unplanned surface contamination between tests.

Optical systems which are used in the low orbits of the US Space Shuttle can be destroyed within some hours due to degrading by atomic oxygen. To investigate these effects the Surface Effects Sample Monitor SESAM was developed. SESAM will be mounted on the ASTRO-SPAS satellite and will carry different sets of optical samples. Different samples will be exposed during the launching, orbit and landing phases. Before and after each exposition phase the surfaces are hermetically sealed. The scattering characteristic of the samples are measured with a modified TIS apparatus, which allows the investigation of bare transparent substrates. Further information on the surface characteristics are collected using Nomarski differential interference contrast microscopy to obtain highly resolving 2D-images.

In this paper, we deal with the laser measurement of surface roughness from spherical optical components (both convex and concave samples) by means of a self-made total integrated scattering apparatus and set out some experimental results. The radius of curvature of the spherical surface can be as much as 50 mm.

The assessment of optical surface flaws requires a standard which objectively reflects their influences upon an optical system and which is also widely accepted. Can the area of flaws, which has been used in most national standards, reflect objectively the performances of optical surface flaws? In this paper, calculation and analysis of the frequency spectrum distributions of typical surface flaws (including scratches and digs) in an optical system have been performed. The relationship between the spectrum intensity distribution and flaw area, depth, and shape has been given. The corresponding experimental results of several 20 micrometers wide scratches with different depths or cross-section shapes have been obtained. From the results obtained above, a novel method which is called Spectrum Energy Function Assessment has been put forward. This method is based on the performance of surface flaws (i.e., spectrum energy) instead of the area of flaws which is widely used in traditional assessment method. The comparison between the two methods has also been described. The results show that the performances of two surface flaws, which are thought to be of the same quality based on the traditional method, are greatly different in the same optical system obtained from the experimental results.

Infrared (IR) Bi-directional Reflectance Distribution Function (BRDF) curves from bare super-polished hot isostatic pressed (HIP) beryllium (Be) and chemical vapor deposited (CVD) silicon carbide (SiC) mirrors are much higher and have drastically different slopes than those predicted from either visible (VIS) BRDF or surtace profilometry data. The end result is that current state-of-the-art HIP Be and CVD SiC will not meet some low scatter requirements. This paper presents data showing that this anomalous IR scattering effect can be easily "covered-up" by coating the bare super-polished substrates.

Our understanding of the relationship between optical surlace topography and scattering behavior has improved in recent years to the point where the scattering characteristics of optical systems are routinely controlled by placing specifications upon the root-mean--square (rms) surface roughness. Bare polished hot isostatic pressed (HIP) beryllium (Be) has consistently failed to obey the established topographic surface scatter models. Surface scatter in the form of the Bidirectional Reflectance Distribution Function (BRDF) from bare beryllium mirrors using several test wavelengths will be presented in this paper. This experimental data indicates that non-topographic scattering effects are significant for bare beryllium mirrors. An empirical "effective" surface power spectral density (PSD) function will be developed that can be used to predict the scattering behavior of bare polished HIP Be from topographic rms surface roughness data.

Bi-directional Reflection Distribution Function measurements were performed as a function of cryogenic temperature for various substrates. Substrates investigated include HIPed and sputtered beryllium produced from different powders and by various manufacturing and polishing processes. In some samples investigated, the BRDF at 10.6 microns increased by a factor of 2 to 5 during cooling from 300 to 30 Kelvin. On repeated temperature cycling the change in BRDF appeared to be totally elastic. The cryo-scatter effect does not occur for all types of beryllium.

To understand the origin of anomalous scatter it is necessary to consider materials-related features which might be responsible for this anomalous behavior. In this study, a variety of material characteristics of a subset of the beryllium mirrors used in an investigation of anomalous scatter by Stover et. al. (1989) is presented, including the near-surface chemical composition, grain size, surface particulate density, and ultraviolet-visible reflectance values. These material characteristics are compared with the anomalous scatter level reported by Stover for the same mirrors. One mirror (B-85), which exhibits a high level of anomalous scatter, was found to have larger grain size (about 28 microns) and a high density of localized surface porosity, compared to other mirror samples with lower levels of anomalous scatter.

This paper discusses the reflection and scattering from slightly-rough crystal surfaces using first-order vector Fresnel-Kirchhoff diffraction theory. Results are given in the conventional, Stokes- and Mueller-matrix representations for cubic and uniaxial crystals in particular orientations.

Anomalous, or non-topographic, scatter from beryllium optics at IR wavelengths has been recognized as a source of poor performance for space based IR imaging systems. This paper reports the results of experiments designed to discover clues as to the source of anomalous scatter so that it can be eliminated with new manufacturing techniques. Polarization, wavelength dependence and surface pitting were examined. None of these can be confirmed as the primary source of anomalous scatter based on the work completed to date.

This paper discusses the wavelength dependence of angle resolved scattering (ARS) from optical surfaces, which have surface roughness and dielectric permittivity fluctuations. Such surfaces are assumed to be of 'optical quality' in that the rms roughness is much less that the incident wavelength. Of particular interest is scattering from beryllium optical components. Depending upon fabrication methods, it has been found that ARS measurements from beryllium at visible through infrared wavelengths do not scale as wavelength to the inverse fourth power. This is contrary to first-order roughness scattering theory that predicts a Rayleigh-like wavelength dependence. This indicates that other non-trivial sources of scattering are also in effect. The results of this work show that a surface that scatters due to surface roughness and dielectric permittivity perturbations along with a statistical correlation between these sources can yield a non-Rayleigh wavelength dependence. This can explain the anomalies seen in beryllium scattering and other materials.

Producibility of low scatter, HIPed beryllium optics requires the manufacturer to know more than just the surface roughness or BRDF of his parts in work. The limitations of his test apparatus (spatial frequency range, height resolution, steepness of slopes, polarization sensitivity, available wavelengths) require that `overlapping' data be taken. This doesn't just mean that the `same' data needs to be taken on similar instruments. Instead, a collection of both quantitative and qualitative data from different types of analysis equipment must be combined to form a more complete picture of the interactions between the material and the wavelengths of interest. This paper discusses the results from several different tests which (when combined) give the manufacturer enough information to determine whether or not there is any more that can be done in his shop to improve the scatter function. We demonstrate that a variety of objective and subjective testing is necessary to determine the `true' characteristics of uncoated HIPed beryllium mirrors. We show results of testing, give a discussion on the interpretation of the data and demonstrate how it was used to optimize production results.

Many groups today are researching the characteristics of beryllium, in an attempt to find ways of producing high quality (low scatter) stable beryllium optics. This paper discusses a two-part study in which (1) an attempt is being made to determine the best, raw beryllium mixture and preparation, machining and polishing processes, test and analysis methods, and (2) a proposed model for the prediction of scatter from beryllium surfaces (based on a knowledge of surface and subsurface interactions with incident wavelengths) will be refined against empirical data. We discuss design of the experiment, the model, and some of the early results.

The paper reviews the design of an instrument designed to detect and map internal defects found in PMMA plastic. The material is used to manufacture intraocular lenses and must be free of defects larger than about 20 micrometers for a variety of safety, manufacturing and cosmetic reasons. The instrument detects scatter from a diode with a CCD array camera to map defect location. This information is used to avoid manufacturing lenses that will eventually be rejected.

The development and use of an automatic system to measure the absolute retroreflectance of various surfaces without any reference standard is described in this paper. Preliminary results are obtained for several samples: defused SiC plates coated with gold, unworked aluminum sprayed with different thickness of paint and metal plate sprayed with different SiN powder and paint mixtures. The results indicate that a correct choice of the parameters allows complementary information of the retroreflection. The accuracy of the system is confirmed to be from 1% to 9% of measured value (for a CO2 laser source), and higher accuracy may be obtained for other wavelength.

A bidirectional reflectometer designed for obtaining both signature data and optical component scattering information must satisfy a wide range of measurement requirements and sometimes conflicting requirements and criteria. The basic definition of bidirectional measurement is reviewed; the design approaches, measurement geometries, measurement procedures and techniques, recommended for a multipurpose instrument are described and discussed.

Output coupler (OC) and high reflector (HR) thin-film coatings and substrates that are employed in 632.8 nm helium-neon (HeNe) lasers are investigated for optical scatter. The measurement of scatter in this paper is in terms of bidirectional reflectance distribution function and calculated total integrated scatter, or BRDF and CTIS, respectively. Laser output power will be briefly reviewed as a function of total scatter loss from the OC and HR. Increasing amounts of loss (scatter), in individual sets of OC's and HR's, reduces the output power of each laser tube from a theoretical optimum output, for the same amount of loss. Sample optics were measured for BRDF and compared with visual microscope inspection. Statistical analysis of many thousands of coated optics provides insight to controlling causes of scatter in the manufacturing process. Here, the average and standard deviation of BRDF scatter data for many optics provide important data for process control. Various scatter data is presented form OC and HR production runs.

A measurement technique has been developed for mapping the bidirectional reflectance distribution function (BRDF) over the entire surface of a curved sample at fixed angles of incidence and scatter. The instrument used was the Toomay, Mathis & Associates, Inc., Complete Angle Scatter Instrument at the Optical Characterization Laboratory (OCL) in Oak Ridge, Tennessee. Raster scans and maps of the BRDF of flat samples are relatively straightforward with this instrument, however, similar measurements of curved samples are more complex. The inherent problems involved in performing BRDF mapping on a non-flat optic will be discussed and one solution to those problems will be described for an Optics Manufacturing Operations Development and Integration Laboratories (MODIL) parabolic assessment mirror.

A fully automated scatterometer, designed for BRDF measurements in the IR at about 10 micrometers , is described. Basically, it works around a reflecting parabola (464 mm diameter, F/0.25) and permits measurements in and out of the plane of incidence. Optical properties of the parabolic mirror are emphasized by a ray-tracing technique which permits determination of the correct illumination on the sample and detection conditions of scattered light. Advantages and drawbacks of such an instrument are discussed, as well as calibration procedures. As a conclusion, we present experimental results to illustrate the instrument capabilities.

The contamination of high-quality, space-borne optics by particles originating from baffle systems could significantly alter the performance of the optics. To assess this potential problem, the bidirectional reflectance distribution function (BRDF) of a low-scatter beryllium mirror was measured at the 10.6-micron wavelength with the mirror in the 'cleaned' state and after controlled contaminations with aluminum oxide powder up to 100 microns in size. The aluminum oxide powder was used to simulate particles which could be released from a typical baffle material. After contamination, the particle size distribution on the mirror surface was statistically sampled using a scanning electron microscope image analysis technique. The BRDF measurements of the contaminated mirror were compared to Mie scattering theory calculations for subwavelength, wavelength, and superwavelength particles sizes.

Attempting to determine the quality and uniformity of a smooth surface from examination of BRDF raster scans can be a risky business as evidenced by studies of commercially grown and polished GaAs wafers. To demonstrate this, raster scan variations with respect to wafer orientation are presented along with a method for analyzing these variations without the need for making multiple raster scans. Some possible sources of the orientational variations are indicated by the data.

Lawrence Livermore National Laboratory (LLNL) has recently developed a wide-field-of- view (28 degree(s) X 44 degree(s)) camera for use as a star tracker navigational sensor. As for all sensors, stray light rejection performance is critical. Due to the baffle dimensions dictated by the large field angles, the 2-part sunshade/baffle configuration commonly seen on space- born telescopes is impractical. Meeting the required stray light rejection performance (of 10-7 Point Source Transmittance, (PST)) with a 1-part baffle required iterative APART modeling (APART is an industry standard stray light evaluation program), hardware testing, and mechanical design correction. This paper presents a chronology of lens and baffle improvements that resulted in the meeting of the stray light rejection goal outside the solar exclusion angle of the baffle stage. Comparisons with APART analyses are given, and future improvements in mechanical design are discussed. Stray light testing methods and associated experimental difficulties are presented.

Paper coatings are porous structures of pigments, binder and air. This research aims at developing a model for the prediction of the bulk optical properties of coating from its structure and material characteristics. Two theoretical approaches have been studied and compared. In both cases the coating is treated as a structured layer consisting of individual scatterers in a medium, and the pores are regarded as scatters. In the first approach they are assumed to be independent. The second approach takes into account that due to their close position the pores are in fact correlated scatterers. The approximation used limits the consideration to pair correlation only. Comparison of the theories to reflectance measurements on polystyrene pigment coatings have uniform pore sizes in the range from 0.13 to 0.95 micrometers showed that for pore size parameters less than 5 the dependent scattering model provides a good prediction for the reflectance, but for larger pore size parameters the reflection behavior approaches that predicted by the independent scattering theory.

The bidirectional reflectance distribution function (BRDF) of both white and black, flown and unflown Space Shuttle thermal protection system tiles has been measured in the 2.0-5.5 micron range. The BRDFs of flown and unflown white tiles were nearly identical, and typical of a good diffuse reflector (with nearly constant BRDF as a function of reflectance angle. The BRDF of the black tile was modified as a result of experiencing Space Shuttle flight; spectral BRDF in this case became a monotonic decreasing function of radiation wavelength.

Because the thermal history of an isothermal metal sphere with an emissive coating can be accurately modelled, such a reference sphere is a suitable calibration object for a space-based IR sensor. To achieve high quality calibration, the uncertainty in the sphere's material parameters must be constrained by relating calibration requirements to design tolerances in the sphere parameters. A methodology for doing this will be presented and applied to an orbiting reference sphere. For clarity, the approach will be illustrated with a gray-body model of the sphere thermal behavior, but results for a non-gray-body sphere will also be given. Sources of uncertainty in the sphere signature will be identified and estimated. In particular, earth flux scatters from the sphere and contaminates the sphere's thermal signature. While the scattered earth flux constitutes a small fraction of a highly emissive sphere's IR signal, it will be shown that the uncertainty in the scattered flux is a significant fraction of the uncertainty in the signal. In sunlight the uncertainty in scattered earthflux reduces the total uncertainty by cancelling other error terms, but in darkness the total uncertainty can be increased by scattering effects.

Several new materials are available for glazing applications, many of which require careful optical characterization, especially with regards to light scattering. Measuring scattering requires special equipment and is inherently difficult. An integrating sphere can be used for the total and diffuse components but great care must be taken in interpreting the instrument readings. Angular resolved scattering measurements are necessary for a complete characterization, and this is difficult for low levels of scattering. In this paper, measurements on electrically switchable NCAP materials and thick panes of aerogel are reported. The NCAP films switch reversibly from a translucent, scattering state to a transparent, clear state with the application of an ac-voltage. Airglass has a porous SiO2 structure with a refractive index n equals 1.04 and a very low heat transfer coefficient. Integrated scattering measurements were performed in the wavelength range 300 to 2500 nm on a Beckman 5240 spectrophotometer equipped with a 198851 integrating sphere. In this instrument we can measure the total and diffuse components of the reflectance or transmittance separately. The angular distribution of the scattered light was measured in a scatterometer, which can perform scattering measurements in the wavelength range 400-1100 nm in both transmittance and reflectance mode with variable angle of incidence.

This paper reports the experimental results on using Langmuir—Blodgett (LB) films as spacer layers between molecules and metal surface to study electromagnetic enhancement mechanisms of Surface Enhanced Raman Scattering (SERS), and to measure the change of SERS intensity with spacer layer thickness. In pyridine + KCL/LB films/Ag films configuration, the experimental results indicate SERS effect exists when the spacer layer thickness is 5 nm, and eventually it becomes unobservable with 15 nm thickness. The experiments supported the idea that the enhancement arises from an electromagnetic mechanism,

Light scattering measurements can be used for determining roughness as well as volume structure parameters of optical thin films. A method of structural analysis is outlined, which allows a quantitative estimation of roughness parameters, mean columnar diameter, packing density, and the evolutionary exponent, respectively.