Several aspects of extracting scene object information directly from the associated autocorrelation (or spectrum modulus) data arrays are investigated. Emphasis is on the particular scenario in which the scene can be modelled by a small set of dispersed objects with associated position, size, shape, and brightness parameters. These parameters may completely define a scene, they may contain the information of interest in a more complex scene, or they may merely constitute a reasonable first approximation to an arbitrary scene. A typical two step approach to estimating such parameters is to first use phase retrieval/image reconstruction techniques to estimate an associated image, and then apply pattern recognition techniques to extract the important information from it. The work described here focusses on eliminating the image recovery step and estimating parameter values directly from the autocorrelation. This task naturally separates into several distinct sub-problems, the first of which is extracting the significant features from the autocorrelation. This is a pattern recognition problem with special consideration to the unique features of autocorrelations. From these feature positions, the number of objects in the scene and their relative positions are next deduced, and finally, the individual object sizes, shapes and brightnesses are extracted. Optional further analysis is described in which the object parameter estimates are further refined by seeking a Maximum Likelihood estimate with regard to the data array. Alternatively, the initial estimates could be used to generate a trial image for an iterative phase retrieval procedure to reconstruct the full scene. Since the trial estimate already contains the major features of the scene, convergence to the correct solution should be both faster and better assured. The phase retrieval problem has been well studied and is not investigated here. For each of these sub-problems, the logical or mathematical development of the solution is presented, implementing computer algorithms are described, and theoretical and practical limitations are discussed.

This dissertation considers multi-wave interactions in bulk semiconductors. Non-equilibrium Green's functions are used to derive an appropriate set of equations describing the interaction of a light field with a semiconductor. Many-body effects lead to the screening of the Coulomb potential in these equations, as well as, carrier-carrier scattering. The carrier-carrier scattering is studied within the context of the carrier Boltzmann equation which contains the dynamically screened Coulomb potential. The relation between carrier scattering and optical dephasing is also made. The carrier scattering rates are then used in the equations describing a two beam pump-probe experiment. The resulting equations are solved numerically for both passive and active systems, and effects such as spectral hole burning, coherent light scattering, and light induced band splitting are studied. Considering three CW beams (pump and two probes) allows for the study of four-wave mixing in semiconductors. By considering CW fields, the semiconductor may be treated in the quasi-equilibrium approximation allowing for greater detail in the treatment of the light field. The resulting probe absorption yields asymmetries which result in an asymmetric four-wave mixing spectrum. The four-wave mixing spectrum is then used to study the amount of squeezing in the light field exiting the cavity in a four-wave mixing experiment.

This dissertation investigates experimentally as well as theoretically several all-optical switching configurations in a rocking filter fiber. It is shown experimentally that the rocking filter can be used as an intensity dependent polarization switch at the resonant wavelength. Up to 70% of the input power consisting of 30 ps pulses could be switched between the orthogonal polarization axes of a 2 long fiber. Stimulated Raman scattering and pulse break-up led to the saturation of the self-switching response. Off-resonant self-switching was shown to yield an enhanced response at shorter wavelengths with lower critical powers than for the resonant case. At longer wavelengths switching was initially inhibited but could still be implemented at powers significantly larger than for shorter wavelengths. This result implies that switching is noreciprocal which is a useful and necessary condition for operating a device as a logic gate. Controlling an intense signal pulse by a weak control pulse through a variable phase delay was successfully demonstrated. In excess of 55% of the input energy could be switched by controlling the phase difference between the two pulses. The phase-sensitive response suffered just as for the self-switching response from Raman scattering and pulse break-up. When combining the nonreciprocity due to wavelength detuning with the phase-sensitive response, logic operation with rocking filters can be implemented. By choosing the proper operating conditions of the device. e.g. input power, phase delay, and wavelength detuning, it is shown that XOR, OR, and AND gates can be realized. The contrast between a logic 0 and 1 state was shown to be as low as 0.3 and 0.7, respectively. Strong pump controlled switching of a weak signal beam produced the best switching characteristics of this dissertation. Up to 90% of the signal pulse could be switched between polarization states. This result was accomplished by using a pump pulse at a different wavelength which had a pulsewidth three times as large than that of the signal pulse.

Multi-photon Doppleron resonances are investigated for the diffraction of atoms by a classical standing-wave optical field in the regime where energy conservation limits the diffraction orders. In this regime, atomic recoil is incorporated into the description, and consideration of the dynamical shift of the resonance condition from the purely kinematic results becomes critical to observing the desired diffraction of the atomic wave function. Special attention is given to the development of a technique for evaluating the resonances which constitutes a significant simplification of the usual degenerate perturbation theory approach for a quantum-mechanical Hamiltonian. Also developed is a specialized off-resonant atom-field interaction which is shown to be maximally efficient at producing high-order diffraction of the atomic wave function. Suggestions for its implementation are given.

For sparsely distributed arrays of optical telescopes operating in phase, large regions of complete attenuation in the modulation transfer function (MTF) result from the spaces between telescopes. These 'holes' in the MTF can be used to advantage to avoid aliasing of high spatial frequency components when the resultant optical image is undersampled by a detector array. The technique for isolating and removing the harmonic components of the sampled image spectrum from the undersampled image has been labeled 'dealiasing'. The result of undersampling the image plane is a reduction in the number of detectors required to image a particular scene by nearly a factor of four. The appropriate conditions for recovering a non-aliased image from the undersampled image focal plane are set by two parameters; the configuration of the pupil, and the detector array sampling rate. The two parameters are coupled, so that the selection of one constrains the allowed values of the other. Errors may be introduced by incorrectly scaling the pupil function to the detector array. Increasing the size of the individual subapertures in the telescope array beyond a prescribed limit also introduces errors that significantly degrade the dealiased result. The latter error is partially mitigated by the associated increase in object information transferred into the image. In this study, an algorithm is formulated to implement the dealiasing technique. The algorithm is demonstrated on point and extended objects using both an image simulation environment and a laboratory breadboard. The robustness of the algorithm to sources of error and the introduction of noise are measured. Quantitative metrics and visual assessment of image quality are used to evaluate the results. The CLEAN algorithm, a method to interpolate between the peaks of the image spectrum of the sparse telescope array is implemented. This enhancement to image quality is useful when demonstrating the concept on detailed, high resolution objects. Finally, a concept for enhancing the image quality of a sparse telescope array is presented. The technique combines imagery of two wavelengths, co-focal on the detector array. One image is of high resolution, undersampled, and dealiased. The second image is Nyquist-sampled but of low resolution. This technique significantly improves on the quality of sparse array imagery obtained at a single wavelength.

Within the framework of interacting-electron models, the optical nonlinearity in π-conjugated polymers is investigated theoretically. A complete microscopic many-body theory is developed for the mechanism of the third order nonlinearity in these materials. The universality and the predictability of the theory are well established. It is found that the bulk of the optical nonlinearity in conjugated polymers is determined by four essential states although all the excited states are involved. It is also found that electron-electron Coulomb interactions in one-dimensional systems play key roles in the nonlinear optical processes and are responsible for the observed spectroscopic features. The theoretical results are compared with experiments in wide variety of materials and excellent agreements between theory and experiments are obtained. With the essential-state theory, various experiments are explained within a single unified theory for the first time, and many controversies are resolved.

Binary optics is a new technology that makes use of the principle of diffraction instead of reflection or refraction to change an incident wavefront. This technology takes advantage of the recent progress in microlithography. There are many new and exciting applications for binary optics, and we can also expect to see the replacement of some conventional optical elements with binary optics. In many ways a binary optic behaves like a diffraction grating with a period that changes continually over the surface of the optic. We find that energy is scattered into different diffraction orders, and there is scattering similar to "grass", "ghosts", "errors of run", "accidental errors of amplitude", and diffuse scattering from surface roughness, just like there is from a diffraction grating. There are vector theories and scalar theories of diffraction. In this dissertation we give the conditions under which the various theories are applicable. We derive a formula for scattering from binary optics with slightly rough surfaces. By comparing this theory to computer simulations of scattering from binary optics we show that the theory can account for random fabrication errors. Formulas are also derived to predict the scattering from systematic errors. The author designed and built an instrument to measure scattering at small angles, and we show that measured scattering from binary optics can be predicted by the theories developed.

Copper phosphate glasses with 40,50 and 60 mole% CuO are melted in air at 1000°C, 1l00°C and 1200°C using quartz or alumina crucibles, and the [Cu²⁺]/[CU(total)] ratio variations with melting time are measured. Glasses are oxidized during melting and reach equilibrium [Cu²⁺] / [CU(total)] ratios which are independent of melting temperature and identical for the 40 and 50 mole% CuO content glasses. Structural considerations rather than a conventional redox reaction seem to determine oxidation-reduction equilibrium of the glass. Thus, the [Cu²⁺] / [CU(total)] ratio in copper phosphate glass can be controlled by changing melting time. In this work, crystallization, optical absorptions, chemical durability, and structure of copper phosphate glass have been investigated depending on the [Cu²⁺] / [CU(total)] ratio in the glass as well as glass composition The crystallization of copper meta phosphate is initiated from the surface and its main crystalline phase is copper metaphosphate (Cu(P0₃)₂), independent of the [Cu²⁺]/[CU(total)] ratio in the glass. However, the crystal morphology, the relative crystallization rates and their temperature dependences are affected by the [Cu²⁺]/[CU(total)] ratio in the glass. The glass transition temperature of the glass increases as the [Cu²⁺] / [CU(total)] ratio is raised. The optical energy gap, E(opt), increases as CuO content and the [Cu²⁺]/[CU(total)] ratio in thE glass are raised. This trend is explained by speculating upon the energy band structure of copper phosphate glass. However, the Urbach energy, ΔE, is affected solely by the [Cu²⁺]/[CU(total)] ratio in the glass. The absorption band of Cu¹⁺ occurs in the vicinity of the absorption tail and is responsible for the increase of ΔE with reducing the [Cu²⁺]/[CU(total)] ratio in the glass. A broad and asymmetric absorption band centered at about 11000 cm⁻¹ is resolved into three component Gaussian absorption bands around 8500 cm⁻¹, 12000 cm⁻¹ and 13250 cm⁻¹. Each of these component absorption bands is assigned to an energy transition between the energy levels in a tetragonally distorted octahedral coordination. The position of the minimum absorption shifts to higher wavenumbers with increasing the [Cu²⁺]/[CU(total)] ratio in the glass regardless of glass composition and is responsible for color changes in the glass. The 55 mole% CuO glass dissolves more readily than either the 40 or 50 mole% CuO glass. As the [Cu²⁺]/[CU(total)] ratio in the glass decreases, the chemical durability improves. Also, the solution pH dependence of the dissolution rate in the oxidized and reduced glasses is found to differ. The structure and Cu²⁺ ion clustering of copper phosphate glass depending on the [Cu²⁺]/[CU(total)] ratio in the glass are investigated by infrared spectra, ESR, SAXS, and TEM.

A major difficulty in determining how well an infrared sensor will perform under actual operating conditions is that testing is often unfeasible due to expense and destructive nature of the tests. These devices are generally tested in a laboratory using a simulator to represent the target. In presenting a scenario to a sensor, there are three distinct areas which must be addressed--the optical system which projects the scene to the detector, the target simulator, and the background simulator. It was determined that the optical system currently in use projected an acceptable scene except in those scenarios where the target was at high altitude against a low radiance background. The mirrors act as room temperature graybody sources which combine to introduce enough unwanted radiation into the projected beam to significantly increase the apparent background temperature. Several single mirror collimators were designed for use in these situations. In designing target models, an important consideration is how the radiation seen by the sensor changes as the target-sensor separation distance changes. At long range, the target appears to the sensor to be a "point" source with radiation that is a sum of all the component radiances of the target. A procedure for representing long-range targets as an equivalent blackbody of equal in-band radiation is described. As the target-sensor distances closes, the individual temperature regions become resolved at the detector. A multi-engine simulator was designed which, when used with the JAWS apparatus, will project the appearance of hot engines and the cooler fuselage. The two most significants improvements designed deal with the problem of projecting the background. The first design is a coldsource which allows projection of colder than room temperature backgrounds to the detector. This is done by limiting the amount of room radiation which can enter the optical train to that necessary to achieve the lower temperature. The next improvement was the use of large sheets of polyethylene of varying thickness to attenuate the radiation projected. By using large sheets, the entire background can be moved, eliminating the frame rate problems inherent in most projector technologies.

The optical measurement of primary mirrors for astronomical telescopes has become increasingly challenging for two reasons. The mirrors, in addition to being larger, are faster and more aspheric in order to shorten the length of the telescope, and the required accuracy of the optical surfaces is more stringent. This dissertation presents improved methods for measuring these mirrors in the laboratory to the required accuracy. The wire test and the scanning pentaprism test, which measure surface slope errors, were designed and run under computer control. The wire test was used to measure the conic constant of a 3.5-m f/1.75 primary mirror to an accuracy of ±0.003 and the scanning pentaprism test measured the conic constant of a 1.8-m f/1 primary to ±0.003. Improvements in these tests were identified that could increase the accuracy significantly. Interferometric optical testing with null correctors is widely used for measuring aspheric surfaces to high accuracy. A system-level analysis of the null test is given. The test is optimized for wavefront accuracy, imaging distortion, and measurement noise from ghost reflections and diffraction. The optical design and analysis of null correctors, including designs for testing 6.5-m f/1.25 and 8.4-m f/1.14 primary mirrors are given. Several new null corrector designs and a method for performing tolerance analysis using structure functions are given. An error in the null corrector, if not detected, would cause the primary mirror to be polished to the wrong shape. (The primary mirrors for the Hubble Space Telescope and the European New Technology Telescope were misshapen because of faulty null correctors.) A new test of null correctors is presented that uses a computer-generated hologram (CGH) to synthesize a perfect primary mirror. When the CGH is measured through the null corrector, it appears as a perfect primary mirror. Apparent surface errors in this measurement can be attributed to errors in the null corrector. A complete error analysis of this test is given. This method has been proven on null correctors for 3.5-m primary mirrors, where it measured errors as small as 5.1 nm rms and confirmed the conic constants to ±0.000078.