This paper is a part of a continuing series of papers (1, 2, 3, 4) that is intended to keep in perspective where the science of stray light has come from, what its present state is, where it is headed, and where it should be headed, as seen from the author's point of view. The latter two items, where it is going to develop and where it should develop, have never found 100% agreement It is our responsibility to both foresee that our technology keeps pace with our enhanced technological capabilities, and to keep pace with the ever more exacting demand of far more detailed representations of tomorrow's optical sensors.

A mathematically rigorous theory of surface scatter that can explain the descrepancies between sur-face statistics determined directly from profiler measurements and those predicted indirectly from scatter measurements is presented. The theory is based on the assumption that the Spectral Density Function (SDF) of the random height variations is the fundamental invariant of the surface and not its Fourier transform, the autocorrelation function. The autocorrelation function and RMS roughness are then calculated from a finite Fourier integral operation whose limits of integration are determined by the measurement bandwidth, and thus, they depend not only on the intrinsic surface statistics but also on the measurement method. A simple form for the SDF that seems to match optical surfaces quite well is assumed. It encompasses surfaces with a definite autocorrelation length and those that have none at all, i.e. fractal surfaces. For the latter, the theory closely resembles that of 1/f noise in electrical systems where the RMS noise (equivalent to the RMS surface roughness) increases with signal bandwidth (equivalent to the surface or incident beam size). An extensive numerical calculation is performed to show that for smooth optical-like surfaces with RMS roughnesses much less than the wavelength, the scatter becomes directly proportional to the SDF if the effective autocorrelation length is set approximately to the surface or incident beam size. This formulation allows one to easily extra-polate the scatter component right down to the specular beam from measurements taken at relatively large angles.

The adhesion force holding particles to a representative engineering optical surface has been investigated. This study addresses the measurement of the removal force from technologically clean surfaces (as opposed to atomically clean). The determination of the adhesion force was by a centrifuge technique, spanning several thousand g's. Particle removal diagnostics were performed optically by measuring residual populations at various g levels. Particles between 10 and 1000 micrometers and various compositions were tested. The results of the experiment have shown that the adhesion forces on similar particles vary immensely. This may help to explain some unexpected transfer phenomena in relatively low-g environments.

Analysis of IR data obtained from several rocketborne sensors in a limb-viewing geometry clearly show off-axis leakage to be the dominant source of radiance at tangent heights above approximately 100 km in all but the strongest of atmospheric emission bands: CO2 v2 (15 gm) and NO (5.3 gm). Telescope leakage model calculations using BRDF values of -2-4 X 10-3 at 10 are in adequate general agreement with much of these flight data and indicate inflight rejection performance 20-40 times worse than model predictions based on pre-flight laboratory measurements. Laboratory measurements on a number of IR telescopes after many months of vacuum storage indicate an average degradation factor of approximately 35 in excellent general agreement with these rocketborne measurements. However, both laboratory and flight measurements indicate degradation factors which are a function of the off-axis angle with a performance at small angles (< 2°) which can be significantly poorer than these average factors, and can result in a major portion of the leakage energy originating from the lower atmosphere. In at least one case, the off-axis rejection performance at small angles was found to be poor enough for non-rejected radiation from the lower atmosphere to become the dominant source of leakage energy in several IR bands. In this example, the clear presence of anomalous atmospheric spectral features in the data for high tangent heights can only be interpreted in this context. Modelling of the spectra from these experiments throughout the IR region requires proper inclusion of the effects of the atmosphere (both in emission and absorption). Results to date indicate that high-resolution line-by-line calculations may be required to fully explain the detailed structure of the spectra from these measurements.

This paper deals with the scattering of electromagnetic waves by rough surfaces. The problem is examined using the Rayleigh hypothesis (i.e. we assume that the field can be represented in the selvedge region by a plane wave expansion). In order to obtain the spectral components of the transmitted field we use the extinction theorem as a boundary condition. This leads to a generalized form of the reduced Rayleigh equations for any case of polarization and for any surface. Then an exact iterative series solution can be obtained. The numerical implementation and properties of convergence of this solution are discussed. Scattering by a grating is considered in details and comparisons with previous rigorous approaches are reported. In addition, it is shown that the case of conical diffraction is easily handled with this formalism.

Wavelength scaling is the process of using scatter data at one wavelength to predict scatter at a different wavelength from the same optical component. The economic advantages of avoiding scatter measurements at all wavelengths of interest are obvious; however, the subtleties involved in making accurate wavelength scaling predictions are not. The requirements of the vector perturbation theory that the optic be a clean, smooth, front surface reflector are not always easily met for all desired wavelengths and materials. Furthermore, the experiments themselves are not trivial because they often involve all the complexities of instrument comparison measurements. This paper presents data for one dimensional (grating like) and two dimensional (polished) reflectors at wavelengths of 1.06, .86, .63, and .49 microns. The results show excellent correlation for both cases.

A lot of confusion exists within the optical community about the definition of Bidirectional Scatter Distribution Function (BSDF), how it should be measured, and how it should be specified. This paper defines BSDF, and then explores the practical factors that limit the measurable upper and lower values of BSDF. The upper value, which is limited by detector aperture size, can take on values exceeding 107 sr-1 in practical cases. The lower limit, sometimes called the electronic noise equivalent BSDF, is determined by the incident power and the detector solid angle as well as the system electronic noise, and needs to be specified with all three values. For systems that are noise limited by the detector, the minimum BSDF can be expressed in terms of the detector parameters (responsivity, noise, active area, and detectivity). The upper and lower limit for several systems has been explored and these results are presented.

An analysis computer program has been designed and written that predicts laser port scatter from the scatter characteristics of each of the port components. First, the scatter data sets from the components are modeled with curve fitting routines. This model is entered into a library for later analysis in predicting port scatter. Ray tracing techniques are utilized along with the component models to predict the angle scatter of the laser port. The resulting predictions are compared to measurements of a fabricated laser port.

Adoption of an aperture shade whose axis is off-set from the optical axis of the SIRTF telescope has been proposed by P.K. Davis of the NASA Ames Research Center. Moving the high side of the asymmetric (truncated) shade away from the optical axis can equalize the barrel baffle lengths illuminated by the aperture shade and reduce the dimensions of both the barrel baffle and aperture shade. The straylight performance of this revised design is compared to that of the SIRTF baseline design, which uses a co-axial asymmetric aperture shade. Both designs incorporate a back focal distance of 100cm and Davis's specifications for the primary and secondary cone baffles. Adoption of an off-set aperture shade for SIRTF would slightly degrade its performance, but would not violate any of the straylight requirements. The SIRTF goal of natural background limited observations can be met with known blacks, Class 750 mirror contamination and either aperture shade configuration.

An appraisal of the optical properties of a satellite metrological sensor, Advanced Very High Resolution Radiometer (AVHRR), is made for a proposed orbital trajectory which results in an unexpected solar straylight path; i.e., bypassing the external scan mirror and Cassegrain telescope, and proceeding directly to the inner conical baffle. This straylight path is directed onto an internal beamsplitter which, in turn, is directly observed by the detectors. Stray light analysis, as a goal, seeks to remove or minimize the influence of such critical scattering paths. This particular straylight path is evaluated using a standard optical ray tracing program, ACCOS V, and represents an uncommon approach to straylight analysis. To determine the expected sensor degradation, a calculation of the expected attenuation coefficient of scattered sunlight in the AVHRR sensor is estimated based on this important critical scattering path.

This paper describes the results of straylight measurements performed on the infrared telescope of the Infrared Background Signature Survey (IBSS) experiment. Included is a description of the sensor and the test facility used to measure straylight performance and surface scatter characteristics at different levels of integration. The measurements are compared with straylight analysis results using the APART/PADE computer code 1. In general good agreement between measurements and analyses were obtained.

The design of vane structure on baffle surfaces is analyzed using the APART/PADE stray light software, and the ASAP optical analysis and stray light software. Vane structure is used to block direct propagation paths from the baffle walls to other objects in the system. The depth and angle of the vanes are variables that are evaluated for both a centrally obscured system and an unobscured eccentric pupil design. It is shown that under most scenarios the vane angle and vane depth are not significant parameters once the first order propagation path is blocked. The significance of a degraded diffuse baffle coating on the various systems is also analyzed. The reason for evaluating two different, but commonly used, sensor designs is to highlight that the results are system design dependent. Concerns of contamination of the optical elements by the vane coatings due to contamination shake off, abrasive launch conditions which create contamination particles, or particle blowoff are also evaluated. BRDF versus RMS roughness is tabulated alongside particle size distributions in order to weight the relative significance of the contamination LEVEL, as defined by MIL-STD 1246A on the various designs.

Opto-mechanical systems which are entered into APART for straylight analysis frequently start out as models on CODE-V. A translator has been coded in FORTRAN which will create the commands for a program one input file in APART based on an existing optical model in CODE-V. Examples of successful translations are presented along with their respective input and output data. Even more interesting are optical models which lead to imperfect translations. Although easily corrected, these cases are used to demonstrate almost insurmountable differences between CODE-V and APART in the definition of the tilts and decenters of their optical entities. In view of the small amount of time required to correct translation errors, there is little benefit in trying to create a smarter translator.

Opto-mechanical systems which are entered into APART for straylight analysis frequently start out as models on a mechanical drafting program such as ANVIL. A translator, described elsewhere, has been coded in FORTRAN which will create the commands for a program one input file in APART based on an existing mechanical model in ANVIL. Having established this link, it is desirable to transfer the resulting opto-mechanical model to a personal computer (PC) for the purpose of report generation. Even more interesting is the ability to modify this model on the PC or to originate a completely new model on the PC for uplink to ANVIL. Using the program CADMOVER, we have effected two way transfers between a VAX computer and an Apple MacIntosh computer. Examples of this activity are provided.

A unique scatterometer and radiation facility has been developed to measure the BRDF at l0.6 μmt of mirror samples cooled to 20°K before and after electron beam irradiation. This paper will describe the facility and present some typical data.

An instrument has been designed and constructed that can measure the bidirectional reflectance distribution function (BRDF) of optical samples at 10.6 μm, at temperatures of about 85 K, and at pressures near or below 10-6 torr. The results indicate, for the metal mirrors tested, that there is no measurable change in the scattering properties when the temperature is varied from room temperature to cryogenic temperatures.

The general performance of the FASCAT (Fully Automated Scatterometer) at Martin Marietta shows that the instrument can make rapid, accurate BRDF (Bidirectional Reflectance Distribution Function) and BTDF (Bidirectional Transmittance Distribution Function) measurements of optical surfaces over a range of approximately ten orders of magnitude (103 to 10-7) in BRDF. These measurements can be made for most surfaces even with the detector at the specular angle because of beam attenuation techniques. Helium-neon and CO, lasers are used as sources in conjunction with a reference detector and chopper. The instrument is controlled by an IBM XT computer, which is also used for data storage. The sample's multiple plane positioning and the detector position angle are controlled by high precision computer driven motor drivers. A gold sample is automatically measured before the sample measurement. An innovative menu system makes the operation of the instrument possible with minimal training.

Raster scan measurement of optical light scatter is described. Comparison is made of angle scan and raster scan data taken with a CASITh/RASITh scatterometer. A new technique for very fast surface scans is presented.

An instrument has been developed that uses coherent optical heterodyne detection to measure the magnitude of the bidirectional scattering distribution function (BSDF). This instrument offers large dynamic range, high angular resolution and high sensitivity.

This paper describes a fully automated bidirectional reflectometer and some of its applications. The main features of the apparatus include bidirectional reflectance distribution function (BRDF), and bidirectional transmission distribution function (BTDF) measurements for both cases of polarization (s and p). In our system, three angular parameters are controlled through the sample holder while the fourth is the angular position of the detector which can rotate in a horizontal plane. The linearly polarized light source is a gas discharge laser using CO (5 gm) or CO2 (10.6 μm). In order to control the incident polarization, we use two tilted ZnSe windows mounted on a rotation stepping motor in the laser cavity. It enables us to maintain a fixed state of incident polarization during the sample rotation. As a consequence of the design of the sample holder, the relative position of the sample and incident beam remains perfectly constant while the detector explores the whole space. Data acquisition, motor displacements and further data treatments including graphical display are performed by a microcomputer. Data for some diffusers are presented. Determination of the emissivity of an opaque solid using bidirectional reflectance data is considered.

We research and manufacture an instrument for measuring surface scattering and roughness of supersmooth surfaces. It can be used not only for measuring opaque sample but also for measuring transparent sample.

A prototype instrument has been developed that separates light scattered from transparent optics into light scattered from the surface and light scattered from the bulk. The instrument also calculates the loss coefficient for transparent materials. The method has been combined with a raster scanning technique to produce a sensitive method of detecting component contamination in the presence of surface scatter.

Welcome to the 3rd Bi-annual panel discussion. The second one was held in 1986 and the first one was in 1977. A few introductory remarks may be helpful in order to get started. I expected two more members, Turney Nue and Tom Leonard, to be here. Maybe they'll come in late. There were two other people I wanted to have here, Bill Wolfe and Hal Bennett. Bill Wolfe has to go to an SPIE meeting; he's the president-elect. Hal Bennett is the guest speaker at a remote location here in town.

In-plane scatter at 0.6328 μ m was measured from a master set of four samples at 18 optical scatter measurement facilities in the United States. They included government laboratories, universities and industry. The samples were a diffuse white, diffuse black, Molybdenum mirror (45 A rms) and Aluminum mirror (7 A rms). Measured scatter at an incidence angle of 10 degrees and a forward scatter angle of 35 degrees varied by a factor of 2 on the diffuse white (mean = 0.27/sr) and a factor of 5 on the aluminum mirror (mean = 61 ppm/sr). The variations in measured scatter values are a result of inadequate calibration techniques, incorrect gain factors, incorrect optical filter factors, inadequate instrumentation or, in some cases, incorrect operating practices. As a result of this round robin study, we are now pursuing a "standard test method" for measurement of scatter. We are also putting increased emphasis on in-house scatter measurement capability, training and availability of reference samples.

Bidirectional Reflectance Distribution Function (BRDF) measurements that characterize the optical scattering properties of surfaces are extremely important to the design of space telescopes and instruments. Sophisticated stray light analysis codes such as APART/PADE require a characterization of the scattering function of optical black surfaces at several angles of incidence and a wide range of detector angles. This paper presents BRDF measurements taken at 5 and 30 degrees incidence of both space qualified black surfaces and some new black surfaces designed for future space instrument and telescope use. These surfaces represent the optical black surfaces most commonly specified for space telescope and instrument baffle and vane surfaces. The scatter measurements were made on a newly built visible and infrared scatterometer designed for accurate measurements of the scatter function of black surfaces.

Far-infrared measurements of the specular and diffuse (BRDF) reflectance of two relatively new IR-black coatings, Ames 24E and Martin Infrablack, are described and compared to previous data on an earlier coating, Martin Black. The specular reflectances of the new coatings are conservatively one (Infrablack) and two (Ames 24E) orders of magnitude less than that of Martin Black across the spectral region from 12μm to 500μm. The reduced reflectance of Infrablack is attributed to a large increase in the roughness of its sub-strate prior to anodization. The diffuse reflectance characteristics of the new coatings are strikingly different, both from each other and from Martin Black. Near normal inci-dence and at wavelengths as long as 200μm Martin Black is very specular; Infrablack is specular-diffuse with special features; and Ames 24E is nearly Lambertian. At 12 and 50μm, the BRDF of Infrablack is considerably less than that of Ames 24E, while at 100 and 200μm they are comparable.

Measurement of near specular scatter requires the specular beam to be tightly focused at the detector plane. The instrument signature is compared to the sample data and the separation point found to determine the minimum measurable scatter angle. Problems arise with this technique when curved samples are measured because the sample changes the beam focus location and requires adjustments to the instrument optics. The result is a focused spot of different diameter. This causes problems when comparing to the instrument signature. This issue, and its correction, are discussed in the paper. Converging samples can be measured closer to specular than flats of similar quality, while diverging samples reduce the ability to measure near specular. Data is presented to illustrate this effect.

Scatterometer optics create near specular scatter (instrument signature) that complicates the process of measuring near specular scatter from samples. In order to achieve measurements within one degree of specular, the instrument signature must be compared to (or subtracted from) the scatter measured from the sample/instrument combination. Because sample scatter is independent of instrument signature, high scatter samples will separate from signature closer to specular than low scatter samples. This paper reports the results of an effort to measure down to few thousandths of a degree from specular. The samples are Bragg cells designed for use in RF spectrometers and the data is taken at 0.86 microns. Data from several materials are presented. The inherent physical limitations imposed on such measurements are also discussed.

The design, analysis, and performance of a small-angle scatterometer are presented. A dye-cell Gaussian apodized aperture is utilized to reduce the small-angle diffraction background. After the diffraction background is sufficiently reduced, the system scatter becomes the dominant noise in the instrument beam profile. The geometrical aberrations are graphically shown to have no significant effect on the higher-order diffraction rings of the beam profile, and as such, an aberrational tolerancing approach can be employed in scatterometer design. Small-angle scattered light measurements are presented for two mirror samples. The associated beam profiles and NEBRDF levels are also given.

Martin Black has been heated to 250 °C in air and the scattering properties at 0.6328 and 10.6 micrometers are described. Significant reductions are seen in the bidirectional reflectance distribution function (BRDF) when the sample is at elevated temperatures. These lower values of the scattering function are seen in both visible and infrared measurements. The reductions in the visible BRDF do not remain after the sample was remeasured after cooling overnight in the normal atmosphere present in the scatterometer. The reduction in the infrared BRDF is however still present in measurements made with the sample at ambient temperature after sitting overnight under the same conditions.

This paper describes the development and use of a system to measure the Bidirectional Reflectance Distribution Function (BRDF) of various surfaces. The BRDF measurements will be used in the analysis and design of optical measurement systems, such as laser anemometers. An argon ion laser (514 nm) was the light source. Preliminary results are presented for eight samples: two glossy black paints, two flat black paints, black glass, sand blasted aluminum, unworked aluminum, and a white paint. A BaSO4 white reflectance standard was used as the reference sample throughout the tests. The reflectance characteristics of these surfaces are compared in this paper.

The Space Telescope program is a major NASA undertaking, the purpose of which is to design, build and operate a National Observatory (known as the Hubble Space Telescope) in space. After feasibility and concept studies lasting well over a decade, the design was begun in earnest in the fall of 1977. It is by far the most significant commitment to the science of astronomy since the 5-meter Hale telescope at Mount Palomar. Space Telescope is a 2.4-meter aperture telescope designed to be 50 times as effective in observing faint objects as the largest ground-based telescopes. Its resolution will be at least 10 times as good. It is to be a long lived facility, capable of on-orbit main-tenance and update as well as ground refurbishment, so that it will be providing useful scientific data well into the twenty-first century. Above the distortion and absorption of the Earth's atmosphere, ST will provide, for the first time, high resolution broad spectral coverage of the entire sky. During the Hubble Space Telescope thermal vacuum test, the measured temperature of the primary mirror was lower than expected, and exhibited an axial gradient, which indicated that heat was being lost from the front surface of the mirror. This is most easily explained by an increase in the surface emissivity of the mirror coating. It was proposed that the small amount of particulate contamination known to be on the surface of the mirror might be the cause of an increase in effective emissivity. However, it could be argued that the conductive coupling between the dust particles and the mirror surface would be very small, making their net effect negligible, thermally. The problem, though conceptually simple to analyze, is not amenable to calculation be-cause we know little or nothing about the physical contact between a dust particle and the host surface. The best way to resolve the question is to measure the effective total hemi-spherical emissivity of a mirror surface with measured levels of particulate contamination. A thermal measurement is preferred, because typical optical measurements, which measure normal emissivity by subtracting measured reflectance from unity, will simply measure the effect of particulate obscuration.

One radiation damage mechanism which has caused particular concern for optical baffle materials is particulate blow-oft. This is an important consideration in the design of advanced optical systems, since particles ejected from the surface of a baffle may foul nearby mirrors and provide a significant source of stray light. To assess the extent of the blow-off problem. and in order to design optical systems which are imo,une to this problem, it is important to know number of particles ejected as well as their size, direction, and speed. We have conducted experiments with several commercial baffle materials to illustrate the nature of the problem and provide insight into the physical phenomena involved.

Optics contamination from small Hydrazine, Al fueled solid, and BeH2/HAP rocket plumes is assessed for space based infrared sensors. Model predictions of BDRF and sensor level degradation caused by erosion cratering and gas condensation are presented as a function of area cratered and film thickness. The resulting reduction in sensor level S/N is determined and expressed as a percentage. Results for the case considered indicate mirror pitting is less than 2% of the clear optical area and film flux is 0.01 g/cm2/sec. The result for 250° K optics is a LWIR reduction in S/N of15`)/0 and a MWIR S/N reduction of 8%. This provides for sufficient sensor operation depending on whether the optics is originally sized to account for this degradation. These results are compared to limited laboratory tests.