This paper describes the application of Multiple Beam Shearing Interferometry (MBSI) to beam collimation. To realize MBSI, a plate with two optically flat surfaces, wedged slightly in thickness at an angle of 2.7 arc seconds and coated with silver is used. A high reflecting coating, in our case with 90% reflectance, is necessary to ensure that multiple beams occur resulting in sharp fringes for the transmitted light. Without the high reflecting coating the transmittance level is unsuitable for beam collimation. The small wedge angle is necessary for high accuracy. When a collimated beam is incident on the plate a sharp horizontal fringe line is observed because of the wedge in the plate. When the beam is not collimated the effect of the wedge and shear of the wavefront results in oblique fringe lines. In this paper, theoretical calculations, sensitivity of collimation and experimental results are discussed.

In our profiler, software is the key-point. During the course of reference height of each point of the surface by computation and comparison of modulation M and so we can get surface profile. In the course of computation and comparison of modulation M, the ability of software will decide whether our profiler can be applied into practice. In this paper, we did some research work in software, and provided three- frame, five-frame and digited filer processing ways.

This paper describes a kind of profiler with the above features. It uses the white-light interfering method and adopts the structure of Micelson interferometer. It includes the light bulb used as the light resource, CCD used as the sensor, PZT providing microscopic translating and the computer sampling and processing data with high speed. When the reference-mirror translating, computing and comparing the modulation M of each point of the surface will give their relative surface height. Then we obtain the surface profile of object to be measured.

A new profiler is under development that uses a commercially available electrostatic device normally used for nanoindentation as its primary sensing element. The new profiler can measure the force applied to the sample and the displacement of the stylus tip in both the vertical (Z) and lateral (X) directions, while the tip is moved in the forward (Y) direction. The sensor has a range of 100 microns and a resolution of 0.2 nm in the Z and X directions. The sensor can measure the force between the tip and the sample with a resolution of 100 nN up to 10 mN. The sensor is normally used to measure indentation force and distance simultaneously during nanoindenting. The instrument allows the force between the stylus tip and the sample to be controlled to a predetermined level and records the displacement of the tip as it scans the sample. Windows based software allows the data to be analyzed for roughness and waviness. The sensor can also move the tip in a lateral (X) direction by applying an electrostatic force to the tip. Lateral motion allows scans to be taken parallel to one another within a 100 micron range. The instrument can be operated in this manner to produce a limited 3D scan of the surface. This instrument offers a low cost device capable of high-resolution profilometry and limited 3D scanning.

Power spectral densities (PSDs) were used to characterize a set of surfaces over a wide range of lateral as well as perpendicular dimensions. Twelve 200-mm-diameter Si wafers were prepared and the surface finished ranged from as-ground wafers to epitaxial wafers. The wafer surfaces were then measured with different methods: atomic force microscopy, angle-resolved light scatter, interferometric and stylus profilometries, and capacitance-based wafer thickness gages. The data were converted into 1D PSDs and the curves were plotted as functions of spatial frequencies. The useful frequency range for each method is indicated and the differences in the calculated PSD values in the overlapping region of two or more methods are discussed. The method used to convert 2D PSDs to 1D ones is presented.

The paper reviews the process for comparing PSD's generated by profile measurements and scatter measurements. Rayleigh scatter as a noise source is reviewed and a new polarization based measurement is presented for discrimination between roughness scatter and particulate scatter. AFM results are compared for both a molybdenum mirror and a silicon wafer. The mirror compares well but 2D detrending technique used impacts the results for the wafer.

It is pointed out that the von-Mises distribution can replace the Gaussian distribution for circular or spherical vector fields, i.e. BRDF data obtained from a variety of technical surfaces by stray light measuring or sensing. For the purpose of in line quality control formulae for the parameters corresponding to mean and variance in Gaussian distributions as well as parameter tests and confidence intervals for circular unimodal vector fields will be given. A family of scatter sensors is introduced. Finally, measurement results will be compared to circular statistical inference.

A high-resolution optical displacement sensing system using an optical waveguide is fabricated and characterized. The measurement principle is based on interference between an even and odd modes in the double-mode waveguide. After the light from a laser light source is focused onto the object to be measured, a wavefront gradient in the converged reflected light beam from the object gives a phase asymmetry at the entrance of the double-mode (DM) waveguide. A change of the phase asymmetry due to the displacement of the object along the optical axis is detected as a change of light intensity distribution at the exit of the DM waveguide. Although the optical system is very simple, experimental results using Ti-indiffused LiNbO₃ waveguide device shows a very high resolution less than 1 nm. Next, a compact-type displacement sensor module using silica waveguide is fabricated and shows an identical high resolution. In conclusion, it will be very useful as a built-in component for various kinds of the industrial equipment.

In the x-ray region, the reflection efficiency of a superpolished surface strongly depends on its roughness. This effect may be used to obtain a 2D map of the roughness spatial distribution for flat surfaces with a rms. roughness height of the order of one nanometer. The basic components of such a device are a precision mechanical 1D scanning stage and a temperature stabilized cooled x-ray linear detector array with quantum efficiency at CuK_(alpha ) radiation.

The use of laser zone texture (LZT) is well established in the data storage industry. LZT is a process where a localized pattern of laser melt bumps is put on the surface of a rigid magnetic recording disk in order to reduce the stiction force between a magnetic recording head and that disk in a hard disk drive. Characterization of the height of the laser bumps on a disk is important to the control of the LZT process. One technique well suited for this measurement and control of LZT in a manufacturing environment is laser Doppler vibrometry (LDV). Commercial instruments using this technique allow fast and precise measurement of an average bump height on a given surface. In addition, excellent correlation of the determined height to atomic force microscopy has been found. In this paper, methods developed for measuring laser bump heights with LDV will be presented along with results on the accuracy and precision of the measurement and correlation to other height measurement techniques. Characterization of different bump shapes and heights and limits of the technique in this respect will also be discussed.

In this paper, the new method on evaluating and measuring roughness of optical surface by CCD camera is presented. The microstructure of surface of samples studied is revealed by a replication technique. The electron photomicrographs of one surface replication of each sample is obtained at two magnification. We note that a grained structure is presented due to residual rough grooves from polishing process. The electron photomicrographs of optical surface is scanned and analyzed by CCD camera-microcomputer system. The corresponding surface profiles are displayed at CRT of microcomputer system, with X-axis corresponding the micrograph spatial coordinate and Y-axis corresponding the micrograph density (i.e. roughness of the real surface). The above-mentioned method has both good vertical and lateral resolution (0.1 - 1.0 nm) due to using CCD camera scanner, but a poor lateral one (100 - 1000 nm) by classical interferometer method. The new method has availability for evaluating and testing optical and supersmooth surface.

The detection of surface particles has become important in contamination control over the years. However, the minimum particle size required to be detected has been becoming smaller as IC geometries shrink. Current visible-light detection systems can detect particles down to around 60 nm in polystyrene-latex-equivalent size and so should be adequate for geometries down to around 0.18 micrometers , but a quick glance at the National Technology Roadmap for Semiconductors shows that geometries are expected to become as small as 0.07 micrometers in a little over ten years, requiring the ability to detect particles around 20 nm in diameter. This is beyond the capability of current visible-light scanners, so the Semiconductor Research Corporation has recently commissioned our group to conduct research into the limits of optical defect detection and potential of alternative detection techniques. This research centers on short-wavelength optical systems and scanned electron-beam systems as the most likely candidate technologies for high- speed nanoparticle detection. In this paper we develop a model for the analysis of the performance of hypothetical short-wavelength surface inspection systems and examine the manifold difficulties involved with using those wavelengths. The properties of scattering in the transitional region between the UV and X-ray regimes are also examined.

The paper reviews techniques to calculate differential scattering cross-section of sub-micron surface bound particles from scatter measurement. It discusses both the multiple and single particle measurement approaches to obtaining scatter data and the associated problems with each method. Measured results are compared to a scatter model based on the discrete source method. For spherical particles the model has very close agreement with the measured scatter. For non-spherical particles the agreement is close only for particle diameters smaller than about one fifth wavelength.

Computer simulations of scatter from various surface features are very important for addressing the design issues associated with producing scanners to meet evolving industry needs. This paper deals with mathematical modeling of light scattering by small pits of arbitrary shape on the surface of silicon wafers. The code based on method of moments applied to the volume integral equation, has been developed, and the numerically modeled differential scattering cross- section is presented for various types of pits. Pit shape and orientation are investigated.

The polarization of light scattered into directions out of the plane of incidence is calculated for microroughness, defects, or particles in or near a dielectric film on a substrate. The theories for microroughness and Rayleigh scatter in the presence of dielectric films are reviewed, and a method for calculating the Mueller matrix for scatter from multiple sources is described. The situation of a 1 micrometers thick layer of SiO₂ on a silicon substrate illuminated by a p-polarized 633 nm light at a 60 degree(s) angle of incidence is used to demonstrate the model calculations. The polarization of scattered light is calculated for scatter from roughness at each of the interfaces separately and roughness of both interfaces when those roughnesses are correlated and uncorrelated. The scatter from a single Rayleigh defect is considered as a function of the position of that defect within the dielectric film, and the scatter from a random distribution of Rayleigh defects in the layer is calculated. The capability of distinguishing amongst different scattering sources is discussed.

Increasing levels of metallization, shrinking device geometries, and stringent defect density requirements have led to a continuous focus in the semiconductor manufacturing community to reduce defects generated during metal deposition by PVD techniques. Pareto analysis of in-film defects in currently used interconnect metallization schemes suggest that a considerable portion of the in-film defects (up to 50%) are caused by Unipolar arcing during Aluminum deposition. Due to their unusual molten appearance, these defects are commonly referred to as `Splats'. These defects can be as large as 500 micrometers , and because of their metallic nature have a high probability of causing device failure. Due to their frequency of occurrence and size, these Splats can significantly impact device yield in a manufacturing environment. Systematic investigations have been carried out for the identification and characterization of these in-film defects, using a combination of Tencor Surfscan, Optical and Ultrapointe Microscopy and SEM FIB analysis. This analysis has revealed these Splats result from localized melting or explosions on the target surface, due to Unipolar arcing. This Unipolar arcing can be strongly correlated to presence of undesirable metallurgical attributes such as Alumina inclusions, porosity, oxygen content etc. in the target. The results of this study indicate that by the reduction/elimination of these various undesirable metallurgical attributes in the Aluminum alloy targets a significant improvement in defect generation during sputter deposition of Aluminum films, and hence an improvement in device yield, is possible.

Detecting and quantifying contaminants and crystalline defects on micro-grade silicon wafers is extremely important to ensure the IC device yield. Although Laser Scanning Surface Inspection Systems are widely used for contaminant inspection, visual inspection continues to be used in silicon wafer manufacturing facilities because conventional particle scanners are not capable of identifying and quantifying a variety of material defects such as Epi spikes, ESF, pits, and sliplines. This work investigates a new technique to detect these material defects by combining information from two independent phenomena, light scattering and reflecting. The optical system to study this technique consists of a conventional particle scanner to detect and quantify light scattering events from contaminants on the wafer surface and a Reconvergent specular Detection apparatus. This apparatus is capable of imaging material defects by measuring attenuation of the light beam intensity reflected from the water surface due to diffraction, absorption, and distortion. Epi mounds voids, slip dislocations, and some other common defect features and contamination on silicon wafers are studied using this equipment. The results are confirmed by that of microscope or AFM. This technology provides the solution to the wafer manufacturing industry for full automated wafer inspection and defect classification.

Applications of Zalar^TM rotation and field emission Auger analysis in various magnetic hard disc problems will be presented. These applications include thin layer structure characterization, low concentration contaminant (0.1 - 0.5 atomic percent) detection at deeper interfaces (greater than 1000 angstroms), and small particle (approximately 1000 angstroms) contaminant identification.

We present a method for the rapid characterization of indium tin oxide (ITO) films. The method determines, from a simple optical measurement, the values of the refractive index (n) and extinction coefficient (k) from 190 to 1100 nm, film thickness, and energy band gap. Also we show that the spectra of the extinction coefficient can be correlated to the film's resistivity. This capability allows the determination of values for the resistivity of ITO films from a very fast and simple optical measurement.

A novel approach for thin film thickness and optical constant extraction from spectral reflectance data is presented here. This methodology combines the global minimization abilities of Adaptive Simulated Annealing with the high computational efficiency of Neural Networks to solve complex characterization problems in real time. The optical constants of many thin films such as Polysilicon are a function of the processing conditions and hence the real time measurement of these parameters could possibly be used in real time or run to run process control applications.

New metrology for characterizing the chemically amplified resist is required to meet the stringent demands of DUV lithography. In this paper, we present a systematic methodology to characterize DUV photoresist by decomposing the resist thin-film into different empirical components according to their optical properties. The innovation of this metrology includes three parts: ellipsometry and reflectometry measurement data de-noising; sophisticated dispersion models derived from the Kramers-Kronig relations for optical thin-film component description; and an intelligent simulated annealing algorithm for the optimization computational engine.

The Non-Repeatable Runout (NRRO) of spindle motor is one of the important parameters in high-recording-density hard disk drive. In this paper, NRRO is investigated by using high precision transducers and measurement systems. A software method is provided to solve the synchronous signal problem that is being presently solved by using hardware apparatus in the same case such as optical encoder etc. This system can be applied for the in-process measurement of NRRO for hard disk drive.