In this paper we describe a method to increase the spatial resolution of surface micro-roughness measurements. As the surface specifications for precision optics become more demanding, the metrology instruments must cover a broad spatial frequency range. Generally, multiple instruments are used to cover the full range of the specifications. For example, an interferometer (Fizeau, Michelson, etc.) would be used to test low spatial frequency surface errors, an interferometric microscope (such as a white light interferomenter) would be used for higher spatial frequency errors, and an AFM would be used for even higher spatial frequency errors. For some precision optics, three or more instruments would be necessary. However, an increase in the resolvable spatial frequency bandwidth of a metrology instrument could reduce the number of instruments necessary to characterize the optical surface over the spatial frequency bands defined by the optical specifications. A solution to increase the resolvable spatial frequency bandwidth of micro-roughness measurements will be presented. This will be accomplished by implementing an interferometric microscope and a process called "sub-pixel spatial resolution interferometry" (SSRI) with interlaced stitching. In this process, multiple interferometric measurements are made as the optic under test (or the CCD array) is laterally shifted at sub-pixel increments. The measurements are then combined to construct a measurement with higher spatial resolution. Initial results obtained implementing a similar process used to increase the spatial resolution of measurements made with a commercially available Fizeau interferometer will be presented.

This article deals with a high-precision three-dimensional positioning and measuring machine and its application as a metrological long-range scanning force microscope. At the Institute of Process Measurement and Sensor Technology of the Technische Universitaet Ilmenau an interferometric nanopositioning and nanomeasuring machine has been developed. Which is able to achieve a resolution of less than 0.1 nm over the entire positioning and measurement range of 25 mm x 25 mm x 5 mm and is traceable to the length standard. The Abbe offset-free design in conjunction with a corner mirror as a reference coordinate system provides extraordinary accuracy. The integration of several probe systems and nanotools (AFM, STM, focus sensor, tactile probes) makes the nanopositioning and nanomeasuring machine suitable for various tasks in the micro- and nanotechnologies. Various probe systems have been integrated in the last few years. For example, a commercial piezo tube AFM was integrated and tested. Additionally, interferometeric measurement systems of the nanopositioning and nanomeasuring machine enables the calibration of probe systems. Also in order to achieve the best possible measurement results special probe systems have been developed and tested and are discussed briefly.

Atomic force microscopy (AFM) and optical diffraction microscopy (ODM) are used to measure the profiles of grating grooves with depths much larger than their widths. Gratings with these features are essential in numerous optical devices such as spectrometers, monochromators and for the production of many fibre Bragg gratings. However, measurement of the physical shape is inherently difficult but necessary for the understanding of their function and in order to improve the manufacturing process. After a thorough calibration of an AFM and by tilting the plane of the grating by up to 17° relative to the symmetry axis of the sensing probe we measured accurately and traceably the sidewall angle and the sidewall profile in a non-destructive way. ODM is a new method where the intensity of the optical field diffracted is measured as a function of the frequency and an inverse algorithm is used to reconstruct the surface profile. It is fast, non-destructive, and it gives height and filling degree of a grating very accurately. As example a high aspect ratio grating with period p of 220 nm, depths d of ≈300 nm, and sidewall angles γ of approximately ≈90° and filling degree f of ≈40 % were examined. Standard uncertainties as low as u(d) = 3 nm, u( α) = 0.4° and u(f) = 3.1 % were achieved. Despite the fact that the AFM responds to the physical surface and ODM responds to the optical properties of the material we find that the results are in very good agreement and consistent with (destructive) scanning electron microscopy measurements of the filling degree.

A static and dynamic 3-D surface profilometer with nano-scale measurement resolution was successfully developed using stroboscopic illumination and white-light vertical scanning techniques. Microscopic interferometry is a powerful technique for static and dynamic characterization of micro electromechanical systems (MEMS). As MEMS devices move rapidly towards commercialization, the issue of accurate dynamic characterization has emerged as a major challenge in design and fabrication. In view of this need, an interferometric microscopy based on white-light stroboscopic interferometry using vertical scanning principle was developed to achieve static and dynamic full-field profilometry and characterization of MEMS devices. A micro cantilever beam used in AFM was characterized using the developed instrument to analyze its full-field resonant vibratory behavior. The first five mode resonant vibration can be fully characterized and 3-5 nm of vertical measurement accuracy as well as tens micrometers of vertical measurement range can be achieved. The experimental results were consistent with the theoretical simulation outcomes from ANSYS. Using white-light stroboscopic illumination and white-light vertical scanning techniques, our approach has demonstrated that static and dynamic 3-D nano-scale surface profilometry of MEMS devices with tens-micrometer measurement range and a dynamic bandwidth up to 1MHz resonance frequency can be achieved.

The development of silicon nano-crystals (nc-Si or dots) technologies requires important characterization work, and till now suffers from the lack of non-invasive, in-line metrology to control the growth of dots. This paper reports the validation of a new in-line dots size and density characterization, achieved by coupling light scattering analysis and XRay reflectivity measurements. A set of nc-Si with density ranging from 10¹¹ to 6x10¹¹cm^-2 and size from 9 to 13nm was grown on Al₂O₃ / HfO₂ stacks and used to validate the correlation between light scattering and dot density, and the correlation between X-Ray reflectivity and dot size. This fast (<10 minutes per wafer) in-line protocol gives very encouraging results, being in good agreement with Scanning Electron Microscopy off-line measurements.

The lateral resolution of an interferometer is limited mainly by the design of the optical arrangement as well as the size of the beam stop. For its characterization the MTF^1,2 is not very useful. The height of a structure normal to the surface under test is transferred into a phase of a reflected wavefront. Since imaging mechanisms for intensity and phase are very different, we propose a Height Transfer Function (HTF) to describe the lateral resolution of interferometers. The HTF shows the quotient of the reconstructed and the original height of a sine-modulated surface structure as a function of the spatial frequency. The HTF can be measured with a test sample of varying periodical surface profiles and spacings. Simulations can be made using a combination of geometrical ray tracing and Fourier transformation techniques. Two different layouts of null systems for the test of an asphere are compared. A device to measure the HTF is shown along with results for a variety of different interferometers.

Increasing technological capabilities to produce microelements (for example: microbeams, micromembranes and micromirrors) and their expanding areas of application introduce unprecedented requirements concerning their design and testing. Conventional two beam interferometry is one of the most popular testing methods of microelements that have reflecting surface. However the elements under test may bring additional challenges: their surface finish may be mixed i.e. reflective-diffusive which restrict their analysis by conventional interferometry; their surfaces may have complicated shape or large shape gradients which restrict their testing by means of interferometer with flat reference mirror. In this paper we propose to solve these problems by converting conventional Twyman-Green interferometer into multifunctional measurement platform by introducing different reference surfaces including: mirror (for conventional two beam interferometer); diffuser (for ESPI); Liquid Crystal On Silicon (LCOS) phase spatial light modulator (for active interferometer). Diffuser allows to implement ESPI in the same system configuration. Special software enables to combine the results of measurement by conventional interferometry (mirror-like surface) and ESPI (diffuse surface). LCOS serves as an adaptive reference mirror and phase shifter. The use of such element allows to increase measurement range of the interferometer and simplifies out-of-plane displacement measurement through object wavefront compensation. The applicability of the platform will be shown at the examples of active micromembranes testing in static and dynamic modes of their work.

A series of phase shifting test mask patterns are described to monitor various effects in image-forming optical systems. Five novel classes of test mask patterns have been developed for in-situ characterization of illumination, aberrations, and mask performance in optical lithography. The monitors combine knowledge of optical effects with the topography enabled by state-of-the-art phase shift masks to create patterns that are sensitive to one aspect of projection printing. Each design is believed to be theoretically the most sensitive pattern to the desired effect. A variety of experimental and simulation studies at 193nm wavelength have validated their scientific principles and have helped allow understanding of limitations due to realistic imaging conditions, most notably the electromagnetic interaction with mask topography. All measurements are conducted in the image plane and are advantageous when access to the remainder of the optical system is limited. Although these monitors have been developed for optical lithography, possible applications in other fields of optics are discussed.

The requirements for form or topography measurement become ever-challenging. Desired lateral resolutions are sometimes smaller than one millimeter, whereas the size of the specimen can exceed one meter. To meet these requirements, scanning measurement systems are increasingly applied with sensors being (much) smaller than the specimen. However, due to the presence of quadratic sensor and scanning stage errors, large errors of the reconstructed topographies can emerge. To overcome these problems, a novel scanning measurement system is proposed. The system includes a linear scanning stage, a compact interferometer used for multiple distance measurements and an autocollimator for additional angular measurements. The effect of quadratic sensor and first-order scanning stage errors which the system allows to eliminate is discussed. Topography reconstruction is outlined and first measurements with a demonstrator set-up are described which show that the proposed novel measurement principle works well under real measurement conditions.

Characterizing an aspheric micro lens is critical for understanding the performance and providing feedback to the manufacturing. We describe a method to find the best-fit conic of an aspheric micro lens using a least squares minimization and Monte Carlo analysis. Our analysis is based on scanning white light interferometry measurements, and we compare the standard rapid technique where a single measurement is taken of the apex of the lens to the more time-consuming stitching technique where more surface area is measured. Both are corrected for tip/tilt based on a planar fit to the substrate. Four major parameters and their uncertainties are estimated from the measurement and a chi-square minimization is carried out to determine the best-fit conic constant. The four parameters are the base radius of curvature, the aperture of the lens, the lens center, and the sag of the lens. A probability distribution is chosen for each of the four parameters based on the measurement uncertainties and a Monte Carlo process is used to iterate the minimization process. Eleven measurements were taken and data is also chosen randomly from the group during the Monte Carlo simulation to capture the measurement repeatability. A distribution of best-fit conic constants results, where the mean is a good estimate of the best-fit conic and the distribution width represents the combined measurement uncertainty. We also compare the Monte Carlo process for the stitched data and the not stitched data. Our analysis allows us to analyze the residual surface error in terms of Zernike polynomials and determine uncertainty estimates for each coefficient.

A digital laser microinterferometer with a capability to measure both static and dynamic properties of MEMS (Micro-Electro-Mechanical Systems) and microstructures has been developed. This system is designed to exploit a static and a dynamic measurements of microstructure such as MEMS with either a smooth or a rough surface. The method is based on digital laser Michelson interferometry and digital speckle pattern speckle interferometry (TV - holography) incorporated with optoelectronic devices including a special illumination system, a long distance microscope (LDM), a CCD camera, a high precise phase shifting unit and a signal generator for vibration measurement. The special illumination system can perform both a continuous (for static measurements) and a stroboscopic (for dynamic measurements) illumination by utilizing a Acoustic-Optic-Modulator (AOM). In this paper, the theory and methodology of the digital laser micro-interferometer with the stroboscopic illumination method are described. The usefulness of the micro-interferometer is demonstrated by examples of static and dynamic measurements for different MEMS.

The Geometry Measuring Machine (GEMM) of the National Institute of Standards and Technology (NIST) is a profilometer for free-form surfaces. A profile is reconstructed from local curvature of a test part surface, measured at several locations along a line. For profile measurements of free-form surfaces, methods based on local part curvature sensing have strong appeal. Unlike full-aperture interferometry they do not require customized null optics. The uncertainty of a reconstructed profile is critically dependent upon the uncertainty of the curvature measurement and on curvature sensor positioning. For an instrument of the GEMM type, we evaluate the measurement uncertainties for a curvature sensor based on a small aperture interferometer and then estimate the uncertainty in the reconstructed profile that can be achieved. In addition, profile measurements of a free-form mirror, made with GEMM, are compared with measurements using a long-trace profiler, a coordinate measuring machine, and subaperture-stitching interferometry.

Combined measurements of transmission T, absorption A and total scattering TS revealed the high accuracy of all applied measurement techniques by obtaining a sum T+A+TS+R = (100±0.3)% (R denotes the Fresnel reflection). In order to investigate CaF₂ at high fluences, a variety of samples from high purity excimer grade to research grade was irradiated (80 ... 150 mJ/cm², 2*10⁶...7*10⁶ pulses) and characterized before and after irradiation by total scattering, laser induced fluorescence (LIF) and transmission measurements. Total scattering mappings showed negligible and measurable scattering in excimer grade and some research samples of minor purity, respectively. For the first time to our knowledge, laser induced fluorescence measurements revealed increasing (580nm, 740 nm) as well as decreasing (313 nm, 333 nm) emissions. The small increases of the linear absorption, obtained in all samples by transmission measurements, were used to distinguish high from minor quality material. For high quality samples the linear absorption change scales with NH³ (N: number of pulses), whereas for minor quality research samples a NH²-scaling was found.

We show that only one measurement of the ellipsometric function ρ at one angle of incidence and one wavelength is totally sufficient to determine the optical constant of the filmN₁, its thickness d, and the substrate optical constant N₂. Obviously, it's also sufficient to only characterize the film; determine N₁ and d, and to only characterize the substrate; determine N₂ and d. A genetic algorithm (GA) is presented that is based on a physical condition of the film-substrate system. This GA is used to characterize the system in those three separate cases. We show that by removing the film thickness from the fitness function of the GA, the computational effort to characterize the film is reduced from 20 000 to 69 calculations. And that to characterize an absorbing layer is reduced from 80 000 to 180. An error analysis is presented that shows the GA is resilient to random experimental errors.

A novel configuration for the channeled spectroscopic ellipsometer (CSE) is presented. The channeled spectroscopic ellipsometry is a snapshot method for the spectrally-resolved polarization analysis. In this method, multiple-order retarders are utilized to generate a channeled spectrum carrying information about the wavelength-dependent multiple parameters of polarization of light. This method has a feature that it requires no mechanical or active components for polarization-control, such as a rotating compensator and an electro-optic modulator. In spite of these advantages, however, the previously proposed configuration of the CSE has a drawback that it is susceptible to the ray-direction variation introduced by the angular fluctuation of the ellipsometric sample. To overcome this drawback, an alternative configuration for the CSE has been developed. In this configuration, the multiple-order retarders are inserted between a light source and a sample, so that the measured results are not affected by the fluctuations due to the reflection from the sample. A compact sensing head whose size is 160mm(W)×53mm(H)×30mm(D) was realized using the new configuration, and applied for the snapshot measurement of the SiO₂ films deposited on a Si substrate, with the acquisition time of 20 ms. The measured thicknesses of the SiO₂ films are almost agree with the results from the rotating-compensator ellipsometer. The configuration that has the multiple-order retarders in the polarization-generating section can apply to other spectroscopic polarimeters to remove the influence of the ray-direction fluctuations due to the reflection or transmission from the sample.

This paper discusses an innovation in reflectometry that presents cost and implementation advantages over configurations currently in use. We demonstrate that the semi-continuous wavelength spectrum used by a spectroscopic reflectometer can be replaced with a small set of fixed wavelength optical sources with only a small loss in measurement accuracy. The resulting instrumentation is called a spectrum sampling reflectometer (SSR). Spectrum sampling can be achieved using inexpensive LEDs and/or lasers as the optical sources. Theoretical calculations for instrument accuracy are shown for both LEDs and lasers along with experimental data from a prototype system using LEDs.

In the paper we present method for three-dimensional measurement of birefringence distribution in photonics components. The tool we used in this work is automated digital photoelasticity combined with tomography, namely photoelastic tomography. Till now the photoelasticity procedures have been applied for objects with significant dimensions and therefore the diffraction phenomena could be neglected. In this paper we verify the correctness of this method for measurement of small objects. In order to identify this error we simulate the measurement process starting from integrated retardation determination and ending on 3D distribution evaluation. Computations are performed using full vectorial propagation method based on Maxwell-curl equations (finite difference time domain method FDTD). The results of simulations are compared with actual measurement results derived from physical setup. The result of experimental and numerical analysis allow optimizing the measurement setup and minimizing the errors. The correctness of the photoelastic tomography for analysis of microobjects has been proven on an example of birefringent optical fibers.

With shrinkage of device size, metrology requirements for Critical Dimension (CD), as defined as the ratio of precision of metrology to process tolerance (P/T), must meet the 0.1 (10%) or 0.2 (20%) criterion.[1][2][14] The precision requirement for gate CD at the 90nm node is thus ~ 0.3nm or less with P/T of 10%, which is far beyond what traditional CD metrology can achieve today. At future nodes, this requirement becomes even tougher, even with P/T of 20%. For years, scatterometry has demonstrated its capability to determine CD and cross sectional profile over periodically aligned line and space (i.e. grating) structures with superior precision. However, to gauge the true capability of scatterometry for process monitoring, the concept of Total Measurement Uncertainty (TMU)[11] of scatterometry in reference to CD-SEM and CD-AFM should be implemented since TMU comprehends both precision and accuracy relative to a reference measurement system. The methodology of implementation of TMU has been discussed in a separate article.[1][12][14][15][17] This paper presents a systematic study on TMU of scatterometry for Final Inspect (FI, post-etch) gate CD and profile, and includes a discussion on how the TMU may be further reduced. One potential option is to feed forward film stack information into the profile modeling, which reduces the number of parameters that have to be calculated during the real-time regression of the scatterometry data.

Diffraction angles, diffraction efficiency and polarization of diffraction by a reflection grating were measured by a null ellipsometer at a wavelength of 632.8 nm. The grating is too rough to be measured by a stylus profiler or an interferometric profiler. The measured diffraction angles follow the grating equation very well and can be used to predict the period of a grating. The efficiency for different diffraction orders can be used to predict the surface profile using appropriate models. For a reflection grating with 150 grooves/mm, the measured ψ ranges from 16 to 84^o and ▵ from 96 to 361^o. This wide range of polarization is rarely seen for other kinds of samples. Depolarization is small when the efficiency is high and efficiency is small when the depolarization is large.

A model for low power (optical output power ≤10mW) InGaAlP lasers operating in the 650nm wavelength band is introduced. This model enables the user to predict lifetime of a diode laser under different operating conditions. Statistically meaningful data can be obtained from the model which gives quantitative values for the considerably increased laser lifetime when operating under less stress conditions.

Light scatter, a bothersome source of noise in optical system, can also be used as a sensitive indicator of surface features. Diamond turned optics typically have three types of surface features. In addition to the background near isotropic (two dimensional) roughness found on most polished front surface mirrors, there are two sources of one dimensional roughness. One of these is the expected periodic tool mark structure that can create a diffraction pattern. The other is random one dimensional roughness (parallel to the tool marks) that is caused by chip drag and jitter between tool and sample. This paper discusses experimental methods for separating the effects of these three distinct sources of roughness and gives examples of measurements on several samples. Several sources of confusion associated with rms roughness specifications (which are often substituted for scatter specifications) are also discussed.

We describe an automated three-axis BRDF measurement instrument that can help increase the physical realism of computer graphics images by providing light scattering data for the surfaces within a synthetic scene that is to be rendered. To our knowledge, the instrument is unique in combining wide angular coverage (beyond 85° from the surface normal), dense sampling of the visible wavelength spectrum (1024 samples), and rapid operation (less than ten hours for complete measurement of an isotropic sample). The gonioreflectometer employs a broadband light source and a detector with a diffraction grating and linear diode array. Validation was achieved by comparisons against reference surfaces and other instruments. The accuracy and spectral and angular ranges of the BRDFs are appropriate for computer graphics imagery, while reciprocity and energy conservation are preserved. Measured BRDFs on rough aluminum, metallic silver automotive paint, and a glossy yellow paint are reported, and an example rendered automotive image is included.

Light scattering techniques allow a comprehensive characterization of surfaces and thin film coatings. Driven by the increasing demands on optical components for DUV lithography at 193 nm, a system for ARS and TS measurements at 193 nm and 157 nm has been developed at the Fraunhofer Institute in Jena. The set-up and measurement examples are presented ranging from ARS of low-scattering substrates and dielectric multilayers to the roughness analysis of EUV mirrors. To follow the recent developments of semiconductor industry towards EUV lithography at 13.5 nm, a new EUV scattering measurement system is developed. The current status is reported.

A study has been underway to provide a better scattering model for front surface reflectors with surface height variations that violate the smooth surface limit required for using the Rayleigh-Rice vector perturbation expression. This paper discusses the issues associated with surface characterization and sample measurement. These include the conversion of 1-D profile measurements to the 1-D PSD, the high frequency correction for the PSD and conversion to the 2-D PSD for use in a scattering model. A single comparison to a new rough surface scattering model is presented. The details of the model are not given in this paper.

We present an experimental study of the angular distribution of light scattered from several rough metallic surfaces, which cover a range of roughness conditions. The substrate materials are steel or glass; roughened by bead-blasting, grinding, or etching; and aluminum- coated. The measured surface-roughness statistics are filtered by using a composite roughness model. The raw mechanical roughnesses range from 0.21μm to 2.66μm; the high-frequency small-scale roughnesses range from 0.13μm to 0.86μm. The optical wavelength is 550nm, so that the roughness-to-wavelength ratio is of order one. A BRDF model based on the Kirchhoff approximation is used to establish a relationship between surface-height statistics and the angular distribution of the scattered light. Angular distributions calculated with the BRDF model are fit to the measurements. The best-fit roughness statistics from the BRDF model agree closely with those measured for the high-frequency small-scale roughness component. The latter roughness component, which has the highest surface slopes, is thus the primary contributor to the angular distribution of the reflected light. We show that the Kirchhoff approximation can be applied to rough metallic surfaces that have multiple scales of roughness and near-, but not perfect, Gaussian surface-height distributions.

By applying Mach-Zehnder interferometer to a pair of symmetrical Collett-Wolf source, the temporal average over an ensemble of realizations of the output power spectrum of the interferometer will display a bright coherent line, which shows the temporal coherence behavior of the input beam. Recently we have further studied the spectral coherence from the interference of a Collett-Wolf source and found a large change in the spectrum of the light.

It has been shown not long ago by Emil Wolf ^(1-2), the Young's interference pattern is formed by partially coherent light. The coherent properties of interfering beams from two pinholes after they passed through a moving diffuser are determined by the correlation function of the heights of the diffuser surface and speed with which the diffuser is moving. We report the recent experimental results displaying the Young's interference of light produced by a pair of symmetric Collett-Wolf partially coherent sources. The experimental results are compared with the results of rigorous computer simulations.

We study theoretically the transmission of p- and s-polarized light through a thin, supported, silver or gold film, both of whose surfaces are one-dimensional, periodically corrugated surfaces, as a function of the wavelength of the incident light. The calculations are carried out by the use of reduced Rayleigh equations in the form of a pair of coupled inhomogeneous matrix equations for the amplitudes of the scattered and transmitted Bragg beams. The results show that enhanced transmission occurs for both p- and s-polarized incident light for large-amplitude periodic corrugations of the two surfaces. In p polarization an additional enhancement occurs at the wavelengths of the surface plasmon polaritons supported by the film. Thus, slits that pierce the film are not necessary for enhanced transmission of light through it.

We report experimental results displaying the interference of light produced by a pair of Collett-Wolf beams. These beams are created by a Michelson interferometer, and are either symmetric or antisymmetric with respect to a plane midway between them. In the former case the output radiation from the interferometer in the far-field is a beam with an intensity distribution that displays a narrow bright line at its center that diverges with the distance from the sources much more slowly than the beam itself. In the later case the radiated beam has an intensity distribution with a narrow dark line at its center. These results suggest that the interference of a pair of symmetric Collett-Wolf beams can be used to produce a pseudo-nondiffracting beam. The experimental results are supported by the results of theoretical calculations.

Thick-film technology is a generally known method for the production of high-end electronic circuit units. The corresponding active electric layers are applied by screen-printing today. The inkjet-printing principle may present an alternative for the future. Aim at the Professorship for Data Processing is the development of such a printing process. Experimental basis is a piezo-driven drop-on-demand (DoD) micro feeding system and a colloidal silver ink. Both, feeding system and ink have to be developed. The ink consists of a solvent (dispersion medium), a dispersing additive and silver and glass particles. Mean particle diameters between 0.2 and 2.5 micrometers are under consideration. One main task is to achieve a homogeneous distribution of the particles on the printed surface. Therefore, the surface of single splashed drops has to be analyzed. Analysis is supported by the Laboratory of Measurement and Information Technology. To characterize a single printed drop with a diameter of about 200 μm every particle has to be seen. Therefore the resolution of an atomic force microscope with the working area of a white light interferometer is necessary. Measurements have been carried out with the Nanometre Coordinate Measuring Machine (NCMM) developed at the Laboratory of Measurement and Information Technology. The NCMM is equipped with an atomic force microscope as a probe and has a working volume of 10 mm x 10 mm x 5 mm. The resolution of the interferometers used is 1.24 nm.

To meet the ever increasing requirements for high quality optical surfaces, components and systems, high-sensitive and flexible analysis of optical losses, nano-roughness and defects is necessary. In this paper, we present the set-ups developed at the Fraunhofer Institute in Jena for total and angle resolved scattering measurements from the VUV and VIS up to the IR spectral regions. Examples are presented for light scattering investigations of the texture of ultraprecision mirrors produced by diamond turning. Experiments are described which were carried out to test the possibilities for nano- and micro defect detection using these set-ups. Furthermore, the relationship between the scattering properties of dielectric multilayer mirrors and their interface texture is discussed for highly reflective DUV mirrors.

Analysis of microscopic images for automatic particle detection and extraction is a field of growing interest in many scientific fields such as biology, medicine and physics. In this paper we present a method to analyze microscopic images of semiconductors in order to, in a non-supervised way, obtain the main characteristics of the sample under test: growing regions, grain sizes, dendrite morphology and homogenization. In particular, nanocrystalline semiconductors with dimension less than 100 nm represent a relatively new class of materials. Their short-range structures are essentially the same as bulk semiconductors but their optical and electronic properties are dramatically different. The images are obtained by scanning electron microscopy (SEM) and processed by the computer methods presented. Traditionally these tasks have been performed manually, which is time-consuming and subjective in contrast to our computer analysis. The images acquired are first pre-processed in order to improve the signal-to-noise ratio and therefore the detection rate. Images are filtered by a weighted-median filter, and contrast is enhanced using histogram equalization. Then, images are thresholded using a binarization algorithm in such a way growing regions will be segmented. This segmentation is based on the different grey levels due to different sample height of the growing areas. Next, resulting image is further processed to eliminate the resulting holes and spots of the previous stage, and this image will be used to compute the percentage of such growing areas. Finally, using pattern recognition techniques (contour following and raster to vector transformation), single crystals are extracted to obtain their characteristics.

This paper reviews work undertaken in the last few years aimed at developing low-cost technology to bridge the gap between subjective visual inspection of specular surfaces and costly objective methods. Since all measurement relies on comparison with a standard the approach adopted is that of the comparison microscope where controlled specimen illumination is used in conjunction with imaging procedures designed to achieve maximum sensitivity to the surface height variations expected. As height variations down to nanometres can be measured with a surface clearance of some millimetres the technology has been termed "Far-Field Nanoscopy". Advantage has been taken of the current wide availability of low-cost digital cameras and associated image processing software. It will be shown how the principal surface metrics of form and finish including roughness, waviness and imperfections can be addressed.

In-line digital holography is presented for the particles analysis in 3-D. The presence of zero order wave, conjugate image wave, and defocused images degrade the image quality in in-line digital holography. In this paper we utilise the numerical reconstruction process to minimise these effects. We present a new subtraction method of reconstructed wave fields for particles analysis. The effects of zero order term, conjugate image wave and more importantly the particles outside the plane of interest (defocused images) are simultaneously minimized, and thus improve the contrast of reconstructed amplitude images. This method is useful to slice in-line digital holograms to display only in-focus particles at any particular plane. Furthermore all this is accomplished from a single hologram with no additional pre or post processing required.

Emissivity and transmissivity of a silicon wafer were studied during the growth of thin oxide films from the viewpoint of spectral, directional and polarized characteristics of thermal radiation. Experimental results were mostly coincident with simulated results. By using a simulation model to estimate the optical properties of silicon wafers, a direct relationship was found between the ratio of p- to s-polarized radiance and the polarized emissivity under specific conditions. This relationship was experimentally confirmed at high temperatures (> 900 K). On the basis of these results, the present study proposes a new radiation thermometry technique that can measure the temperature and spectral emissivity of a silicon wafer at a wavelength of 0.9 μm and at moderately high temperatures, irrespective of the variation in emissivity with oxide film thickness.

This paper presents an overview of a photometer developed by CEA to characterize large coated optics (up to 800 x 400 mm²).

In this paper, we describe size-effect of fluorescent microsphere for measuring point-spread function (PSF) in confocal fluorescence microscopy. We present the numerical results for the practically available microsphere size range for measuring PSF, and demonstrate with experiment. Also, the effective PSF is restored with deconvolution technique within an acceptable error. Also the size-effect for measuring phase modulated PSF, which has two vicinal peaks, is described. The numerical and experimental result is also presented.

White light interferometry (WLI) has played an important role in nano- and micro-scale profile measurement technology. To meet the demand of high-accuracy, high-repeatability, and cost-effective measurement, the research activities on WLI and its applications are rapidly in progress. WLI is based on the superposition of waves with different but very close wavelengths to produce beat phenomena (or to generate detectable envelopes in the interferogram) and then to identify the locations of the zero-order interference fringes (or those of the maximum intensities of interference fringes without the optical path difference). The locations reflect the information of three dimensional (3D) surface profiles from the consecutively acquired images in a WLI system. The objective of this paper is to develop a high-accuracy and cost-effective WLI measurement system, especially for the surfaces of micro-mechatronic devices, micro-optical components, semiconductor devices, etc. In our approach, the feasibility of the use of spectral coherence properties to meet the system design requirement is first investigated. Specifically, proper spectral filters are employed to enhance the coherence length of the light source (i.e., that of the filtered light source) to an order of ten micrometers. Then, a Young's double-slit interference experiment with filtered and unfiltered white light sources is conducted to demonstrate the effectiveness of this technique. Also, we adopt a Michelson interferometric configuration as the optical module of the proposed WLI system, for the sake of its simplicity. Experimental results indicate that several inexpensive spectral filters, a lower-grade charge-couple-device (CCD) image sensor, and a PZT (piezoelectric transducer) with a lower movement resolution are merely needed to develop the WLI system, instead of the use of higher-grade optical and optomechanical components. It turns out that the proposed system with high-accuracy measurement performance is more cost-effective than others.

A high resolution Confocal Laser Scanning Microscope (CLSM) with UV / VIS light sources was developed as the first step of multi-functional microscope. The optical system is designed to optimize for both UV and VIS wavelengths. An UV laser is used to achieve higher resolution, and a VIS is for multi functions. A new objective lens specialized for this application was designed and fabricated. Specification of the lens and the optical system is NA:0.95, EFL:2.5mm, WD:1.5mm, Resolution:160nm and achromatic for two wavelengths of UV 325.0nm / VIS 632.8nm. Several specimens were characterized to check the performance of the system. Some optical materials under study were measured for evaluation, and interesting results could be obtained. Multi-functional measurements are being planed as a next step. This system will help the research of nano-structures, photonic-crystals and biology.

Thick resist lithography is a rather complicated process, which involves quite a few nonlinear factors, so surface profile is largely affected by process conditions such as baking, exposure and development parameters. In this paper, the photochemical reaction mechanism of the thick diazonaphthoquinone(DNQ)-Novolak based photoresists is discussed in detail, and then the effect of exposure intensity on the photochemical reaction speed is investigated by using kinetic model. Numerical simulation and experimental results are presented and agree well with each other. Through comparison between the simulated and experimental results for thick DNQ-Novolak based resists, the photochemical reaction speed is obviously affected by the intensity magnitude during exposure, which will lead to the failure of exposure reciprocity law. This phenomenon is caused by the increase in temperature of resists, due to the highly exothermic reaction during exposure. These results are useful for the lithographic process optimization of thick film resists.

We present results for the effective surface impedance tensor (EIT) of polycrystals of metals in a weak uniform magnetic field H. The frequency region corresponds to the region of the local impedance boundary conditions applicability. We suppose that the resistivity tensor rho_ik(H) of single crystal grains out of which the polycrystal is composed, is known up to the terms O(H²). For olycrystals of metals of arbitrary symmetry elements of EIT can be calculated within the same accuracy in $H$, even if the tensor rho_ik(H)is strongly anisotropic. As examples, we write down EIT of polycrystals of (i) cubic metals, (ii) metals with ellipsoidal Fermi surfaces, (iii) metals of tetragonal symmetry whose tensor rho_ik(0) is strongly anisotropic. Although polycrystals are metals isotropic in average, in the presence of a uniform magnetic field the structure of EIT is not the same as the structure of the impedance tensor of an isotropic metal with a spherical Fermi surface. The obtained results are exact in the framework of the approximation used when describing single crystal galvanomagnetic characteristics. They cannot be improved neither in powers of H, nor with respect to the anisotropy of single crystal grains.